[{"data":1,"prerenderedAt":538},["ShallowReactive",2],{"footer-primary":3,"footer-secondary":93,"footer-description":119,"tv-mcp-showcase":121,"tv-mcp-showcase-seasons":132,"tv-mcp-showcase-episodes":147,"sales-reps":286},{"items":4},[5,29,49,69],{"id":6,"title":7,"url":8,"page":8,"children":9},"522e608a-77b0-4333-820d-d4f44be2ade1","Solutions",null,[10,15,20,25],{"id":11,"title":12,"url":8,"page":13},"fcafe85a-a798-4710-9e7a-776fe413aae5","Headless CMS",{"permalink":14},"/solutions/headless-cms",{"id":16,"title":17,"url":8,"page":18},"79972923-93cf-4777-9e32-5c9b0315fc10","Backend-as-a-Service",{"permalink":19},"/solutions/backend-as-a-service",{"id":21,"title":22,"url":8,"page":23},"0fa8d0c1-7b64-4f6f-939d-d7fdb99fc407","Product Information",{"permalink":24},"/solutions/product-information-management",{"id":26,"title":27,"url":28,"page":8},"63946d54-6052-4780-8ff4-91f5a9931dcc","100+ Things to Build","https://directus.io/blog/100-tools-apps-and-platforms-you-can-build-with-directus",{"id":30,"title":31,"url":8,"page":8,"children":32},"8ab4f9b1-f3e2-44d6-919b-011d91fe072f","Resources",[33,37,41,45],{"id":34,"title":35,"url":36,"page":8},"f951fb84-8777-4b84-9e91-996fe9d25483","Documentation","https://docs.directus.io",{"id":38,"title":39,"url":40,"page":8},"366febc7-a538-4c08-a326-e6204957f1e3","Guides","https://docs.directus.io/guides/",{"id":42,"title":43,"url":44,"page":8},"aeb9128e-1c5f-417f-863c-2449416433cd","Community","https://directus.chat",{"id":46,"title":47,"url":48,"page":8},"da1c2ed8-0a77-49b0-a903-49c56cb07de5","Release Notes","https://github.com/directus/directus/releases",{"id":50,"title":51,"url":8,"page":8,"children":52},"d61fae8c-7502-494a-822f-19ecff3d0256","Support",[53,57,61,65],{"id":54,"title":55,"url":56,"page":8},"8c43c781-7ebd-475f-a931-747e293c0a88","Issue Tracker","https://github.com/directus/directus/issues",{"id":58,"title":59,"url":60,"page":8},"d77bb78e-cf7b-4e01-932a-514414ba49d3","Feature Requests","https://github.com/directus/directus/discussions?discussions_q=is:open+sort:top",{"id":62,"title":63,"url":64,"page":8},"4346be2b-2c53-476e-b53b-becacec626a6","Community Chat","https://discord.com/channels/725371605378924594/741317677397704757",{"id":66,"title":67,"url":68,"page":8},"26c115d2-49f7-4edc-935e-d37d427fb89d","Cloud Dashboard","https://directus.cloud",{"id":70,"title":71,"url":8,"page":8,"children":72},"49141403-4f20-44ac-8453-25ace1265812","Organization",[73,78,84,88],{"id":74,"title":75,"url":76,"page":77},"1f36ea92-8a5e-47c8-914c-9822a8b9538a","About","/about",{"permalink":76},{"id":79,"title":80,"url":81,"page":82},"b84bf525-5471-4b14-a93c-225f6c386005","Careers","#",{"permalink":83},"/careers",{"id":85,"title":86,"url":87,"page":8},"86aabc3a-433d-434b-9efa-ad1d34be0a34","Brand Assets","https://drive.google.com/drive/folders/1lBOTba4RaA5ikqOn8Ewo4RYzD0XcymG9?usp=sharing",{"id":89,"title":90,"url":8,"page":91},"8d2fa1e3-198e-4405-81e1-2ceb858bc237","Contact",{"permalink":92},"/contact",{"items":94},[95,101,107,113],{"id":96,"title":97,"url":8,"page":98,"children":100},"8a1b7bfa-429d-4ffc-a650-2a5fdcf356da","Cloud Policies",{"permalink":99},"/cloud-policies",[],{"id":102,"title":103,"url":81,"page":104,"children":106},"bea848ef-828f-4306-8017-6b00ec5d4a0c","License",{"permalink":105},"/bsl",[],{"id":108,"title":109,"url":81,"page":110,"children":112},"4e914f47-4bee-42b7-b445-3119ee4196ef","Terms",{"permalink":111},"/terms",[],{"id":114,"title":115,"url":81,"page":116,"children":118},"ea69eda6-d317-4981-8421-fcabb1826bfd","Privacy",{"permalink":117},"/privacy",[],{"description":120},"\u003Cp>A composable backend to build your Headless CMS, BaaS, and more. \u003C/p>",{"id":122,"title":123,"logo":124,"cover":125,"tile":126,"announcement_text":127,"description":128,"slug":129,"one_liner":130,"card_text":127,"status":131,"sort":8},"5938933b-ddff-4283-ae4c-e0460e85112a","MCP Showcase","585f431e-5771-4610-9676-e8563022f6ac","23596e6e-aa87-475f-b1cf-8c8b8b0f7926","ec743a55-2bce-414f-aa55-2c3aa3b32b6b","New","Catch up on our MCP Showcase event - ​where we demoed the MCP live, featured what our community had already built, and kicked off an MCP hackathon with great prizes.","mcp-showcase","Catch our MCP Showcase from 20th November on demand.","published",[133],{"id":134,"number":135,"show":122,"year":136,"episodes":137},"49af1154-dd24-49c1-a1fd-fb2da0a4484b",1,"2025",[138,139,140,141,142,143,144,145,146],"f1b56b45-9398-41c1-becd-1d08d15e593d","629c73cb-a886-4db1-9e45-a604316de145","3bf8ea55-3ff8-4dc8-937b-093ca713b8a9","f1c9f6ee-8760-4719-8834-e4da28f9b3eb","7df39746-4a7e-4916-ba9a-71c1fa291944","437dea6b-a229-449c-9437-3f6620c2898d","95ed81c1-d6f4-4f12-817c-ca7aaeb9ac85","3ac5e6eb-111e-4bcf-b2ab-66e70712b3a8","1848dd2c-dbff-4361-9571-eb64a629c46d",[148,164,181,195,210,226,243,258,272],{"id":138,"slug":149,"vimeo_id":150,"description":151,"tile":152,"length":153,"resources":8,"people":8,"episode_number":135,"published":154,"title":155,"video_transcript_html":156,"video_transcript_text":157,"content":8,"seo":158,"status":131,"episode_people":159,"recommendations":161,"season":162},"mcp-101","1138979993","Join Bryant as he talks through the details of the Directus MCP.","305db0ef-78f3-4f8c-b646-c2edd912da3c",3,"2025-11-20","MCP 101: What It Actually Does","\u003Cp>Speaker 0: Hey, everyone. Brian here. I gotta say, I'm super excited for this event. But let's get the intros out of the way, and then we'll dive in. If you're new to Directus, it's a headless CMS and back end that lets you build and manage any kind of data or content.\u003C/p>\u003Cp>You design your own data model or connect your existing SQL database. Directus will generate production ready APIs for you, and you get a complete admin app to manage everything on top of that. So let's move on to the main course. Right? This is what you're all here for.\u003C/p>\u003Cp>What is MCP? Why is it so freaking hyped? Let's talk about it. Model context protocol or MCP is basically just a way for your AI assistance like Claude or ChatGPT or your IDE like Cursor or Versus Code to connect directly to your tools and data sources. Instead of copying and pasting between ChatGPT and your CMS, the AI, it can talk to Directus directly, and we've built MCP support right into Directus Core.\u003C/p>\u003Cp>You enable it in the AI settings, generate an access token for a new user or existing user, connect your favorite AI tools, and you're done. That's it. No middleware, no extra servers to manage, no nothing to deploy. Sounds great. Right?\u003C/p>\u003Cp>But what does this actually do, Brian? Well, hey. That's where it gets interesting. Right? There are multiple different use cases depending on your role.\u003C/p>\u003Cp>So for content teams, you can import blog posts straight from Google Docs with proper field mapping and not having to waste an hour reformatting the copy. You You can organize hundreds of images that are all named image underscore whatever. Add metadata to the entire asset library without clicking through each file individually, spending a ton of time. For developers, you can quickly prototype data models just by describing what you need. Hey, AI.\u003C/p>\u003Cp>Build me a CRM with organizations, contacts, and deals. It will go through and create the collections, set up the relationships for you, the whole thing. Or you can modify your schema on the fly. Hey, I AI. I need a reading time field for all my blog posts and calculate it for my existing articles, set up an automation for that.\u003C/p>\u003Cp>Your AI tools work through the existing Directus permissions. So the beautiful thing about this is it can only do what you've allowed it to do. That keeps everything nice and secure, and everything shows up in your activity log. It's basically your normal direct us workflow, just faster, which brings us to what we're showing today. In this showcase, we're going to walk through some real examples, content creation, data modeling, automation, the stuff that people are actually using this for in production.\u003C/p>\u003Cp>We're gonna show you what works well, what to watch out for, and how to get the most out of combining Directus with AI. And you'll probably also hear a lot of rabbit puns. So with that, down the rabbit hole we go.\u003C/p>","Hey, everyone. Brian here. I gotta say, I'm super excited for this event. But let's get the intros out of the way, and then we'll dive in. If you're new to Directus, it's a headless CMS and back end that lets you build and manage any kind of data or content. You design your own data model or connect your existing SQL database. Directus will generate production ready APIs for you, and you get a complete admin app to manage everything on top of that. So let's move on to the main course. Right? This is what you're all here for. What is MCP? Why is it so freaking hyped? Let's talk about it. Model context protocol or MCP is basically just a way for your AI assistance like Claude or ChatGPT or your IDE like Cursor or Versus Code to connect directly to your tools and data sources. Instead of copying and pasting between ChatGPT and your CMS, the AI, it can talk to Directus directly, and we've built MCP support right into Directus Core. You enable it in the AI settings, generate an access token for a new user or existing user, connect your favorite AI tools, and you're done. That's it. No middleware, no extra servers to manage, no nothing to deploy. Sounds great. Right? But what does this actually do, Brian? Well, hey. That's where it gets interesting. Right? There are multiple different use cases depending on your role. So for content teams, you can import blog posts straight from Google Docs with proper field mapping and not having to waste an hour reformatting the copy. You You can organize hundreds of images that are all named image underscore whatever. Add metadata to the entire asset library without clicking through each file individually, spending a ton of time. For developers, you can quickly prototype data models just by describing what you need. Hey, AI. Build me a CRM with organizations, contacts, and deals. It will go through and create the collections, set up the relationships for you, the whole thing. Or you can modify your schema on the fly. Hey, I AI. I need a reading time field for all my blog posts and calculate it for my existing articles, set up an automation for that. Your AI tools work through the existing Directus permissions. So the beautiful thing about this is it can only do what you've allowed it to do. That keeps everything nice and secure, and everything shows up in your activity log. It's basically your normal direct us workflow, just faster, which brings us to what we're showing today. In this showcase, we're going to walk through some real examples, content creation, data modeling, automation, the stuff that people are actually using this for in production. We're gonna show you what works well, what to watch out for, and how to get the most out of combining Directus with AI. And you'll probably also hear a lot of rabbit puns. So with that, down the rabbit hole we go.","7f40dfe7-1149-467b-85e9-20e38e08c07e",[160],"c8f4c795-4fb6-46ca-b7d2-00c019ffb026",[],{"id":134,"number":135,"show":122,"year":136,"episodes":163},[138,139,140,141,142,143,144,145,146],{"id":139,"slug":165,"vimeo_id":166,"description":167,"tile":168,"length":169,"resources":8,"people":8,"episode_number":170,"published":154,"title":171,"video_transcript_html":172,"video_transcript_text":173,"content":8,"seo":174,"status":131,"episode_people":175,"recommendations":178,"season":179},"setting-up-mcp","1138989005","Join Bryant and Lindsey as they show you how to get started with the MCP.","1420d770-1835-4ddb-85c4-f293d92e775b",5,2,"Setting up MCP: Start Building Today","\u003Cp>Speaker 0: Alright. I'm here with Lindsay from our team. And in this segment, we're gonna show you how to connect the Directus Native MCP to your LLM or your AI tools of choice. Lindsey, do you want to walk us through the process?\u003C/p>\u003Cp>Speaker 1: Yeah. So the first step is enabling it. So you'll wanna go to your Directus instance, go into settings, and then navigate to the AI section. And here's where you'll set up your MCP server. So you'll just click enable.\u003C/p>\u003Cp>If you want the prompt collection, you can turn it on here, or if you already have it, you can find it and tell us what it is. And then make sure you click that save at the top. Decide if you wanna allow deleting as well.\u003C/p>\u003Cp>Speaker 0: And what is our next step after we've enabled the MCP?\u003C/p>\u003Cp>Speaker 1: Yes. So now this is important. You'll wanna set up a new user so that you can set up a clean token with the access permissions you need for your MCP. So we're gonna give it administrator right now, but you might want it to give it slightly tighter permissions if you wanna keep it from doing everything. We're gonna do this because we're gonna have it messing with the schema.\u003C/p>\u003Cp>Speaker 0: Alright. So now we've got our token. We are ready to connect to, our tools of choice. What are we gonna start with first, Lindsay?\u003C/p>\u003Cp>Speaker 1: So let's do cursor. So that's one of our easiest ones. If you go to our docs, we actually have a one click button to get you started there. So find the cursor section in our docs, scroll down, and click a button. It's gonna open up your cursor, and you'll just fill in the settings here.\u003C/p>\u003Cp>So you'll wanna put in your URL for your instance, and then this is where you'll grab that token that you created from your user.\u003C/p>\u003Cp>Speaker 0: Alright. And I think we'll just hit install.\u003C/p>\u003Cp>Speaker 1: And there you go. You're good\u003C/p>\u003Cp>Speaker 0: to go. And we we can see we are connected, and then we are ready to chat. And as you can see, the Directus MCP is connected to cursor, and we are actually picking up the schema. Alright. So that is how you connect the native MCP to Cursor.\u003C/p>\u003Cp>Super easy to get started. Make sure you check it out.\u003C/p>\u003Cp>Speaker 1: But what about marketers? They typically use chat g p t five or something similar.\u003C/p>\u003Cp>Speaker 0: That's a a great question, Lizzie. So let me show you just how easy it is to connect to ChatGPT. Again, once you've got that token, all you need to do is log in to ChatGPT. In the bottom, you'll go to your specific settings. You'll look for apps and connectors, and you have to be in developer mode, which sounds scary.\u003C/p>\u003Cp>It's really just, a beta feature that they have for protection, and security. So once you activate developer mode, then you can go in and add your own MCP. They're called custom connectors. So we'll just add Directus. We'll give it our URL here.\u003C/p>\u003Cp>There we go. Simplecmsstarter.directus.app/mcp. And then we're going to add a question mark access underscore token, and then you're going to paste that same access token that you have. We're going to choose no authentication, and then we just hit create. There we go.\u003C/p>\u003Cp>And now I should be able to chat. So we'll just ask it what our Directus schema looks like. It should pick up the available tools through the MCP, and it'd be able to understand my full schema inside Directus. Again, super simple to connect with ChatGPT, and happy MCP ing.\u003C/p>\u003Cp>Speaker 1: If you have any feedback for us or any questions, head over to community.directus.io and let us know.\u003C/p>","Alright. I'm here with Lindsay from our team. And in this segment, we're gonna show you how to connect the Directus Native MCP to your LLM or your AI tools of choice. Lindsey, do you want to walk us through the process? Yeah. So the first step is enabling it. So you'll wanna go to your Directus instance, go into settings, and then navigate to the AI section. And here's where you'll set up your MCP server. So you'll just click enable. If you want the prompt collection, you can turn it on here, or if you already have it, you can find it and tell us what it is. And then make sure you click that save at the top. Decide if you wanna allow deleting as well. And what is our next step after we've enabled the MCP? Yes. So now this is important. You'll wanna set up a new user so that you can set up a clean token with the access permissions you need for your MCP. So we're gonna give it administrator right now, but you might want it to give it slightly tighter permissions if you wanna keep it from doing everything. We're gonna do this because we're gonna have it messing with the schema. Alright. So now we've got our token. We are ready to connect to, our tools of choice. What are we gonna start with first, Lindsay? So let's do cursor. So that's one of our easiest ones. If you go to our docs, we actually have a one click button to get you started there. So find the cursor section in our docs, scroll down, and click a button. It's gonna open up your cursor, and you'll just fill in the settings here. So you'll wanna put in your URL for your instance, and then this is where you'll grab that token that you created from your user. Alright. And I think we'll just hit install. And there you go. You're good to go. And we we can see we are connected, and then we are ready to chat. And as you can see, the Directus MCP is connected to cursor, and we are actually picking up the schema. Alright. So that is how you connect the native MCP to Cursor. Super easy to get started. Make sure you check it out. But what about marketers? They typically use chat g p t five or something similar. That's a a great question, Lizzie. So let me show you just how easy it is to connect to ChatGPT. Again, once you've got that token, all you need to do is log in to ChatGPT. In the bottom, you'll go to your specific settings. You'll look for apps and connectors, and you have to be in developer mode, which sounds scary. It's really just, a beta feature that they have for protection, and security. So once you activate developer mode, then you can go in and add your own MCP. They're called custom connectors. So we'll just add Directus. We'll give it our URL here. There we go. Simplecmsstarter.directus.app/mcp. And then we're going to add a question mark access underscore token, and then you're going to paste that same access token that you have. We're going to choose no authentication, and then we just hit create. There we go. And now I should be able to chat. So we'll just ask it what our Directus schema looks like. It should pick up the available tools through the MCP, and it'd be able to understand my full schema inside Directus. Again, super simple to connect with ChatGPT, and happy MCP ing. If you have any feedback for us or any questions, head over to community.directus.io and let us know.","8478d1a2-218b-43df-9764-f61c177d8dc3",[176,177],"5e4704dd-c8d2-4efe-8162-0b029f34259a","5ed24299-b070-4b50-91f3-c32623700d24",[],{"id":134,"number":135,"show":122,"year":136,"episodes":180},[138,139,140,141,142,143,144,145,146],{"id":140,"slug":182,"vimeo_id":183,"description":184,"tile":185,"length":170,"resources":8,"people":8,"episode_number":153,"published":154,"title":186,"video_transcript_html":187,"video_transcript_text":188,"content":8,"seo":189,"status":131,"episode_people":190,"recommendations":192,"season":193},"ai-in-directus","1138986971","Rijk shared how we're approaching AI at Directus.","b2888cf3-877e-439a-a153-62e152f40260","Our Perspective on AI in Directus","\u003Cp>Speaker 0: Hello. It's Reik van Zonthe, cofounder and CTO here at Directus. Very excited with this MCP launch. Very excited to see what you're gonna integrate it with and what new use cases you might find. For this segment, I was asked to go in a little bit deeper on the philosophy of artificial intelligence here at Directus, how we treat it right now, how we're planning on using it now and in the future, and what it means for our end users.\u003C/p>\u003Cp>So over time, over the last months and years, we've seen a lot of different directions that AI can take, a lot of possibilities that were unlocked. Some of them very exciting and real. Some of them a little bit silver bullet magic. At the end of the day, for us, it's all about helping our users achieve their goals in the most efficient and easy to use way as possible. We're not here to just replace a bunch of tasks that a person might have to do.\u003C/p>\u003Cp>We're here to help you out get to your results faster. With that in mind, we're really treating this more as a collaboration tool rather than an automation tool. Christina will do a little segment on that later, so keep watching. We found that LLMs are pretty great at helping unblock users in tasks they're taking, but do wanna make sure that we offer those LLM tools when and where they make sense. So for our very first deep dive here into AI, we started with MCP as the focus area.\u003C/p>\u003Cp>That allows our users to be unblocked when using third party tools like Claw, JetGPT, Recursor, or any other tool that supports MCP, and allow them to use their content, their data within Directus from those tools directly. When it comes to AI use, we obviously have to make sure that we have some very strict control about what an LM is allowed to read, what it is allowed to do with your data. And in general, this this fits in neatly with DirectUs' permission system because we, of course, through policies in our RBAC system, we have some very strict access control in place. The same applies for your MCP, usage. In general, we're implementing AI here as a carefully considered approach with that real user value rather than gimmicky for the sake of it.\u003C/p>\u003Cp>I'm incredibly excited to see what you will integrate our MCP endpoint with. Please do let me know as well. As this technology progresses, we're seeing a lot more interesting use cases, things we might have never even thought of on how to use it. So please let us know all of your thoughts, feelings, concerns, cool experiments, and everything else on community.directus.ao. And keep your eyes peeled because we have a lot more cool AI stuff coming in the near future.\u003C/p>\u003Cp>Thank you.\u003C/p>","Hello. It's Reik van Zonthe, cofounder and CTO here at Directus. Very excited with this MCP launch. Very excited to see what you're gonna integrate it with and what new use cases you might find. For this segment, I was asked to go in a little bit deeper on the philosophy of artificial intelligence here at Directus, how we treat it right now, how we're planning on using it now and in the future, and what it means for our end users. So over time, over the last months and years, we've seen a lot of different directions that AI can take, a lot of possibilities that were unlocked. Some of them very exciting and real. Some of them a little bit silver bullet magic. At the end of the day, for us, it's all about helping our users achieve their goals in the most efficient and easy to use way as possible. We're not here to just replace a bunch of tasks that a person might have to do. We're here to help you out get to your results faster. With that in mind, we're really treating this more as a collaboration tool rather than an automation tool. Christina will do a little segment on that later, so keep watching. We found that LLMs are pretty great at helping unblock users in tasks they're taking, but do wanna make sure that we offer those LLM tools when and where they make sense. So for our very first deep dive here into AI, we started with MCP as the focus area. That allows our users to be unblocked when using third party tools like Claw, JetGPT, Recursor, or any other tool that supports MCP, and allow them to use their content, their data within Directus from those tools directly. When it comes to AI use, we obviously have to make sure that we have some very strict control about what an LM is allowed to read, what it is allowed to do with your data. And in general, this this fits in neatly with DirectUs' permission system because we, of course, through policies in our RBAC system, we have some very strict access control in place. The same applies for your MCP, usage. In general, we're implementing AI here as a carefully considered approach with that real user value rather than gimmicky for the sake of it. I'm incredibly excited to see what you will integrate our MCP endpoint with. Please do let me know as well. As this technology progresses, we're seeing a lot more interesting use cases, things we might have never even thought of on how to use it. So please let us know all of your thoughts, feelings, concerns, cool experiments, and everything else on community.directus.ao. And keep your eyes peeled because we have a lot more cool AI stuff coming in the near future. Thank you.","019e56b8-a87f-4f7a-9d65-e362c204d375",[191],"40d5c649-a40e-4a79-950d-505597ed81dc",[],{"id":134,"number":135,"show":122,"year":136,"episodes":194},[138,139,140,141,142,143,144,145,146],{"id":141,"slug":196,"vimeo_id":197,"description":198,"tile":199,"length":153,"resources":8,"people":8,"episode_number":200,"published":154,"title":201,"video_transcript_html":202,"video_transcript_text":203,"content":8,"seo":204,"status":131,"episode_people":205,"recommendations":207,"season":208},"collaborative-cms-mcp","1138969642","Christina talks through what we mean at Directus by 'collaborative CMS' and how the MCP is part of the story.","3407ff52-ee9a-47c3-966b-e3ebd4933606",4,"What is a Collaborative CMS?","\u003Cp>Speaker 0: Hi. I'm Christina. I'm director of product marketing at Directus. And I'm here to clear up one question. So what does collaborative CMS mean?\u003C/p>\u003Cp>The problem with CMS is has always been that any given product is geared either to developers who are building a CMS for their colleagues or their clients, or for people who are the nontechnical end users of the CMS that the devs are building. So marketing tends to think mainly in campaigns, narratives, and speed to market. Engineering tends to think mainly in data structures, validation rules, and system integrity. So this isn't a skills gap between them. It's something else.\u003C/p>\u003Cp>The result is that developer focused platforms give you power and flexibility, but they're opaque to non platforms give you power and flexibility, but they're opaque to non technical users. User friendly platforms make content editing accessible, but they constrain what the developers can build. Every time you build a CMS, you have to pick your poison and choose an architectural trade off. What we're exploring though with our new MCP tool at Directus and what we've implemented natively is bridging that divide. And here's what it looks like.\u003C/p>\u003Cp>A developer can build and change collections through the natural language interface of the MCP so that they can ship faster. You can add and delete fields, insert data into your collections, build automations like flows, edit existing collections, and create multiple new collections while still configuring all the kinds of fields that are added, like making them non nullable or adding conditions, setting up validation rules, and of course, creating relationships. Meanwhile, a marketer can describe what they need in their own language. They can use the MCP to completely change the brand tone of voice and pieces of writing, edit length, shorten content, generate SEO metadata, and alter slugs. Of course, they can update items and blog posts at scale and like you'd expect, they can create new content in bulk.\u003C/p>\u003Cp>So whichever user is tasked with doing something like importing the entire product catalog along with pricing tiers and regional variations, the AI understands this request and the underlying data schema creating properly structured entries with the right relationships and validation that developers have previously established. So this is the bridge between what developers and marketers traditionally do independently. The MCP allows either person to handle that middle work, and that speeds up both teams. And this is what we mean by collaborative CMS. And the MCP becomes a natural language interface for people to do all kinds of work together.\u003C/p>\u003Cp>Both technical and non technical users working in the same system, speaking in their own language without needing to translate their needs into another context or language. And what we're seeing so far is development teams setting up robust data models once. Marketing teams operating within those guardrails at their natural velocity. There's no tickets, no waiting, no translation errors, and it's an end to all the rework. So could natural language interfaces remove communication friction between teams?\u003C/p>\u003Cp>We think so. We're seeing it happen with our MCP integration. Companies are already using it to change how their teams work together. We're excited for you to try it. Let us know what you think.\u003C/p>","Hi. I'm Christina. I'm director of product marketing at Directus. And I'm here to clear up one question. So what does collaborative CMS mean? The problem with CMS is has always been that any given product is geared either to developers who are building a CMS for their colleagues or their clients, or for people who are the nontechnical end users of the CMS that the devs are building. So marketing tends to think mainly in campaigns, narratives, and speed to market. Engineering tends to think mainly in data structures, validation rules, and system integrity. So this isn't a skills gap between them. It's something else. The result is that developer focused platforms give you power and flexibility, but they're opaque to non platforms give you power and flexibility, but they're opaque to non technical users. User friendly platforms make content editing accessible, but they constrain what the developers can build. Every time you build a CMS, you have to pick your poison and choose an architectural trade off. What we're exploring though with our new MCP tool at Directus and what we've implemented natively is bridging that divide. And here's what it looks like. A developer can build and change collections through the natural language interface of the MCP so that they can ship faster. You can add and delete fields, insert data into your collections, build automations like flows, edit existing collections, and create multiple new collections while still configuring all the kinds of fields that are added, like making them non nullable or adding conditions, setting up validation rules, and of course, creating relationships. Meanwhile, a marketer can describe what they need in their own language. They can use the MCP to completely change the brand tone of voice and pieces of writing, edit length, shorten content, generate SEO metadata, and alter slugs. Of course, they can update items and blog posts at scale and like you'd expect, they can create new content in bulk. So whichever user is tasked with doing something like importing the entire product catalog along with pricing tiers and regional variations, the AI understands this request and the underlying data schema creating properly structured entries with the right relationships and validation that developers have previously established. So this is the bridge between what developers and marketers traditionally do independently. The MCP allows either person to handle that middle work, and that speeds up both teams. And this is what we mean by collaborative CMS. And the MCP becomes a natural language interface for people to do all kinds of work together. Both technical and non technical users working in the same system, speaking in their own language without needing to translate their needs into another context or language. And what we're seeing so far is development teams setting up robust data models once. Marketing teams operating within those guardrails at their natural velocity. There's no tickets, no waiting, no translation errors, and it's an end to all the rework. So could natural language interfaces remove communication friction between teams? We think so. We're seeing it happen with our MCP integration. Companies are already using it to change how their teams work together. We're excited for you to try it. Let us know what you think.","bad4717c-1b52-4ec0-a2e7-75ba94a65b42",[206],"8b786101-b7a5-498f-b45b-bf9450894079",[],{"id":134,"number":135,"show":122,"year":136,"episodes":209},[138,139,140,141,142,143,144,145,146],{"id":142,"slug":211,"vimeo_id":212,"description":213,"tile":214,"length":215,"resources":8,"people":8,"episode_number":169,"published":154,"title":216,"video_transcript_html":217,"video_transcript_text":218,"content":8,"seo":219,"status":131,"episode_people":220,"recommendations":223,"season":224},"10-mins-tiktok-clone","1138963281","Bryant and Ben try to build the project of Ben's dreams using the MCP, in 10 minutes or less!","95397771-2625-415b-a579-1335736d00ce",12,"Building in 10 Minutes or Less: TikTok Clone","\u003Cp>Speaker 0: Alright. Welcome back to another episode of one app in ten minutes. This is a extra special season or, extra special show. Normally, this is 100 apps, one hundred hours. Today, we're doing everything in ten minutes.\u003C/p>\u003Cp>That comes with a lot of hype and A\u003C/p>\u003Cp>Speaker 1: lot of pressure.\u003C/p>\u003Cp>Speaker 0: This extra special episode. We've got mystery app with Ben. So mystery app on two levels. One, Ben, our CEO, had no idea he was being pulled into this. And two, I've got no idea what he wants to build, but we're gonna see if we could do it in ten minutes.\u003C/p>\u003Cp>Ben, welcome to the show.\u003C/p>\u003Cp>Speaker 1: Thank you, Brian. I'm very excited to mysteriously be here.\u003C/p>\u003Cp>Speaker 0: Yes. Alright. So the name of the game today, ten minutes to plan and build, no more, no less. You might wanna be thinking about what app we're gonna be building now. I'm gonna tell everybody else how we're making this possible.\u003C/p>\u003Cp>So I've got a blank instance of Directus set up. I've also got our Directus MCP server connected to Claude AI, and that is the the secret sauce that we're gonna be using today to try and go much faster. So, alright, what do you what do you wanna build today, Ben?\u003C/p>\u003Cp>Speaker 1: My my first thought is, like, a content like a like a TikTok.\u003C/p>\u003Cp>Speaker 0: TikTok.\u003C/p>\u003Cp>Speaker 1: It's just, like, an easy Look at that. TikTok clone. TikTok clone.\u003C/p>\u003Cp>Speaker 0: With the owner. This CEO. Alright. 10 on the clock. Alright.\u003C/p>\u003Cp>Next up, you know how this goes. I know you've seen the show. We're gonna start the clock. We're gonna talk about what functionality we need, and then we're gonna go. Alright.\u003C/p>\u003Cp>Clock is running. Alright, Ben. What do we need out of, a TikTok clone?\u003C/p>\u003Cp>Speaker 1: Oof. We need to we need a feed, like a FYP, like a little just getting through getting through the content, getting through the videos. Need to be able to interact with those, so I guess give it a thumbs up. So, like, a little a favorites toggle.\u003C/p>\u003Cp>Speaker 0: Like a likes. Mhmm.\u003C/p>\u003Cp>Speaker 1: And then comments. I don't know what else is\u003C/p>\u003Cp>Speaker 0: Comments and probably like a post. Right? The Yeah.\u003C/p>\u003Cp>Speaker 1: I guess the content itself is like a video, the author. So I guess they should all be attached relationally to an author, and then a description, a time stamp, posted time stamp.\u003C/p>\u003Cp>Speaker 0: Okay. Alright. Let's let's give this a shot. Right? We're gonna just copy paste this.\u003C/p>\u003Cp>We're gonna go to our trusty AI. You know, excited to be able to maybe separate fact and hype here. So, let's build a TikTok clone inside Directus. We want the following functionality. Personalized feeds, posts with likes and comments, and let's just use what?\u003C/p>\u003Cp>The direct Us users collection. Yeah. Mhmm. Let's use I was absolutely driving Reich crazy on another episode with typos. So let's use\u003C/p>\u003Cp>Speaker 1: Is it apostrophe missing a typo?\u003C/p>\u003Cp>Speaker 0: Yeah. There we go. Let's use direct users as the user's collection and reuse off there. Alright. We're gonna hit go, and we're gonna see what happens.\u003C/p>\u003Cp>Speaker 1: I can I can vouch for the fact that I had no clue what this is gonna be, so you definitely didn't? So there's\u003C/p>\u003Cp>Speaker 0: no Oh. Oh, no. Is this a oh, boy. Alright. We're just gonna add to the difficulty level here.\u003C/p>\u003Cp>It looks like I wasn't originally connected to the MCP. So we'll just refresh and go again. How we doing on time? Seven minutes forty seven seconds. Alright.\u003C/p>\u003Cp>We\u003C/p>\u003Cp>Speaker 1: burned through a lot on that on that hour.\u003C/p>\u003Cp>Speaker 0: No pressure. Right? If if one hour wasn't enough drama for you, let's do it all in ten minutes. So My water's\u003C/p>\u003Cp>Speaker 1: empty. We're still over there.\u003C/p>\u003Cp>Speaker 0: Alright. So let's see. We've got the direct assist assistant system prompt here. It's fetching the schema. Obviously, this is a blank instance, and then it's looking for the directest users collection.\u003C/p>\u003Cp>Alright. So now we can see that the AI is trying to go in and and create these collections inside direct us. And it looks like we've got posts, we've got likes, we've got comments, follows, hashtags for post discovery.\u003C/p>\u003Cp>Speaker 1: Relational\u003C/p>\u003Cp>Speaker 0: hashtags. And it looks like it's got a junction table for posts and hashtags. Mhmm. Cool. And, I I don't think you've had a ton of experience with the MCP yet other than our like, the video updates.\u003C/p>\u003Cp>Only my content. So the it it approaches this very systematically, and then it will go in and create the collections. And it looks like it has only created one collection at a time. Alright. So this is new behavior to me.\u003C/p>\u003Cp>Previously, it's done all the collections and then the relations, but we we should be getting the same effect.\u003C/p>\u003Cp>Speaker 1: Like that time excuse there, Brian.\u003C/p>\u003Cp>Speaker 0: Yeah. Yeah. Yeah.\u003C/p>\u003Cp>Speaker 1: Yeah. Absolutely. Yeah. Yeah.\u003C/p>\u003Cp>Speaker 0: Alright. So Oh, wow. So we've already got the likes. We could see the posts. Got a caption, a video file, a thumbnail, user ID.\u003C/p>\u003Cp>Is this all does this line up with what you would expect out of out of this bin?\u003C/p>\u003Cp>Speaker 1: Yeah. I mean, my design brain is is, like, thinking of, like, a little bit of form layout, but I'm impressed that it's doing, like, half in full width, and it's kinda doing a really good first pass.\u003C/p>\u003Cp>Speaker 0: Here's our awesome video.\u003C/p>\u003Cp>Speaker 1: And, I don't have any. So let's see. That was my first concern is, like, you're not gonna have any, like, MP fours at the ready.\u003C/p>\u003Cp>Speaker 0: We've got tons. Recdus MCP assets. What do we got? Blog post workflow. We could test the the TUS upload protocol at the same time.\u003C/p>\u003Cp>Speaker 1: So, Meta, are you uploading an m c p p video for the m c p video?\u003C/p>\u003Cp>Speaker 0: I was trying to. It looks like, some I'm getting some errors. So that's part of the fun of this. We just keep refreshing and and seeing more content or more data model show up. So we got post.\u003C/p>\u003Cp>Let's try this again. I'm gonna link this to my admin user. Here's our awesome video, And let's upload a video from device, see what it got. MCP assets. Do I have a video that is actually small enough?\u003C/p>\u003Cp>It looks like 30 megabytes is all I've got here.\u003C/p>\u003Cp>Speaker 1: It feels like that that can work. Except I'm getting a issue.\u003C/p>\u003Cp>Speaker 0: Oh. Oh. No stress. Alright. Let's just try an image for now.\u003C/p>\u003Cp>Oh, I've done something wrong with the images as well.\u003C/p>\u003Cp>Speaker 1: I think TikTok has images, doesn't it?\u003C/p>\u003Cp>Speaker 0: It it it does. It looks like something has gone wrong. I had reversed relationships. Okay. So now it's filling in the other relationships.\u003C/p>\u003Cp>And, unfortunately, I don't think it's gonna allow us to do this without a file, but we could solve for that really quickly.\u003C/p>\u003Cp>Speaker 1: I know. I'm googling sample m p four file. I like cityscapes and all that on the mind I've seen. I think it is something that I've done here. Thirty seconds, 1.5 megs.\u003C/p>\u003Cp>1.5 megs. Schema.\u003C/p>\u003Cp>Speaker 0: It's like Add to proceed.\u003C/p>\u003Cp>Speaker 1: Does this not know we are on a\u003C/p>\u003Cp>Speaker 0: okay. So I spotted the issue, and that's a good point to bring up that this is really cool functionality to be able to prototype this quickly, but it's not a it's not a silver bullet. It looks like the MCP had tried to tag the wrong folder that did not exist inside the instance, and that's where the issue was coming from. So we could try this again. Here's our test post.\u003C/p>\u003Cp>Let's add admin user. Just going to I\u003C/p>\u003Cp>Speaker 1: Slacked your URL if you wanna try doing it from URL. Okay.\u003C/p>\u003Cp>Speaker 0: Alright. We'll try it from the URL. There we go. We've got the video file uploaded. We can set this to publish.\u003C/p>\u003Cp>We will save our post. And now that the MCP has done its thing, we got two minutes and forty seconds Infinite. To, let's just create create some test posts and sample content for each collection. Cool. Alright.\u003C/p>\u003Cp>In the meantime, Ben, what is great about Directus and how fast you could prototype like this?\u003C/p>\u003Cp>Speaker 1: I mean, this is this is insane. Like, with all the things that we're encountering here I mean, even just the boilerplate stuff of getting the prompt right, trying to find just a sample video file. Like, I you just I feel like you can just, like, mess around. And the worst case scenario, you just command z it and then start again, or at least get, like, a good 80% of the way there.\u003C/p>\u003Cp>Speaker 0: We got one minute and thirty six seconds. I think one of the the coolest things to showcase here is that we have a ready to go API. Right? And not just, I need, like, one API. We've got all the APIs for what we need on the back end of this thing.\u003C/p>\u003Cp>Speaker 1: I love that it's doing all the colors and iconography too. Like, that to me is it just takes so much time. Like, oh, okay. What what, like, material symbol icon should I use for this or that? Like, that plus the colors and all the form layout.\u003C/p>\u003Cp>Like, that's\u003C/p>\u003Cp>Speaker 0: Yeah. So let's go in and create a a like here. Here's our post. Cool. They created should be populated automatically.\u003C/p>\u003Cp>We got follows. We got comments. We can add a comment to the post. Test comment. Alright.\u003C/p>\u003Cp>User ID. Commenting on my own\u003C/p>\u003Cp>Speaker 1: I got a user. Who said who said test comment?\u003C/p>\u003Cp>Speaker 0: There we go. Alright. So thirty seconds left. I I feel like, you know, we didn't obviously do anything on the front end here, but now we we've got these APIs with a a little bit more time. We could probably slap together a front end, and have something up and running in under an hour, which, you know, the namesake of the show is is insane to me that you could do so much with DirectUs in an hour.\u003C/p>\u003Cp>But being able to get here in ten minutes, The entire back end is,\u003C/p>\u003Cp>Speaker 1: like, the the schema. We've got, like, everything sort of configured the right way. We got the full API, REST, GraphQL.\u003C/p>\u003Cp>Speaker 0: That's it. Ten minutes is up. Well, Ben, any kind of final words on Directus and AI before we we wrap this episode up?\u003C/p>\u003Cp>Speaker 1: I mean, it's just it's super exciting. Like, it's the confluence of two big things. Like, AI is just, like, variable, ephemeral, like, crazy new emergent technology with all this potential, but it's kind of a little errant with all the craziness that can get thrown into the mix. And then you've got this, like, super stable, reliable, resilient back end platform of directors. I think, like, at the intersection of those two things is is, as Jonathan would say, there there's magic there.\u003C/p>\u003Cp>Yeah. No. This is this is awesome.\u003C/p>\u003Cp>Speaker 0: And with that, we will end this episode. Thanks for joining us, and stay tuned for more.\u003C/p>\u003Cp>Speaker 1: Thank you, Brian.\u003C/p>","Alright. Welcome back to another episode of one app in ten minutes. This is a extra special season or, extra special show. Normally, this is 100 apps, one hundred hours. Today, we're doing everything in ten minutes. That comes with a lot of hype and A lot of pressure. This extra special episode. We've got mystery app with Ben. So mystery app on two levels. One, Ben, our CEO, had no idea he was being pulled into this. And two, I've got no idea what he wants to build, but we're gonna see if we could do it in ten minutes. Ben, welcome to the show. Thank you, Brian. I'm very excited to mysteriously be here. Yes. Alright. So the name of the game today, ten minutes to plan and build, no more, no less. You might wanna be thinking about what app we're gonna be building now. I'm gonna tell everybody else how we're making this possible. So I've got a blank instance of Directus set up. I've also got our Directus MCP server connected to Claude AI, and that is the the secret sauce that we're gonna be using today to try and go much faster. So, alright, what do you what do you wanna build today, Ben? My my first thought is, like, a content like a like a TikTok. TikTok. It's just, like, an easy Look at that. TikTok clone. TikTok clone. With the owner. This CEO. Alright. 10 on the clock. Alright. Next up, you know how this goes. I know you've seen the show. We're gonna start the clock. We're gonna talk about what functionality we need, and then we're gonna go. Alright. Clock is running. Alright, Ben. What do we need out of, a TikTok clone? Oof. We need to we need a feed, like a FYP, like a little just getting through getting through the content, getting through the videos. Need to be able to interact with those, so I guess give it a thumbs up. So, like, a little a favorites toggle. Like a likes. Mhmm. And then comments. I don't know what else is Comments and probably like a post. Right? The Yeah. I guess the content itself is like a video, the author. So I guess they should all be attached relationally to an author, and then a description, a time stamp, posted time stamp. Okay. Alright. Let's let's give this a shot. Right? We're gonna just copy paste this. We're gonna go to our trusty AI. You know, excited to be able to maybe separate fact and hype here. So, let's build a TikTok clone inside Directus. We want the following functionality. Personalized feeds, posts with likes and comments, and let's just use what? The direct Us users collection. Yeah. Mhmm. Let's use I was absolutely driving Reich crazy on another episode with typos. So let's use Is it apostrophe missing a typo? Yeah. There we go. Let's use direct users as the user's collection and reuse off there. Alright. We're gonna hit go, and we're gonna see what happens. I can I can vouch for the fact that I had no clue what this is gonna be, so you definitely didn't? So there's no Oh. Oh, no. Is this a oh, boy. Alright. We're just gonna add to the difficulty level here. It looks like I wasn't originally connected to the MCP. So we'll just refresh and go again. How we doing on time? Seven minutes forty seven seconds. Alright. We burned through a lot on that on that hour. No pressure. Right? If if one hour wasn't enough drama for you, let's do it all in ten minutes. So My water's empty. We're still over there. Alright. So let's see. We've got the direct assist assistant system prompt here. It's fetching the schema. Obviously, this is a blank instance, and then it's looking for the directest users collection. Alright. So now we can see that the AI is trying to go in and and create these collections inside direct us. And it looks like we've got posts, we've got likes, we've got comments, follows, hashtags for post discovery. Relational hashtags. And it looks like it's got a junction table for posts and hashtags. Mhmm. Cool. And, I I don't think you've had a ton of experience with the MCP yet other than our like, the video updates. Only my content. So the it it approaches this very systematically, and then it will go in and create the collections. And it looks like it has only created one collection at a time. Alright. So this is new behavior to me. Previously, it's done all the collections and then the relations, but we we should be getting the same effect. Like that time excuse there, Brian. Yeah. Yeah. Yeah. Yeah. Absolutely. Yeah. Yeah. Alright. So Oh, wow. So we've already got the likes. We could see the posts. Got a caption, a video file, a thumbnail, user ID. Is this all does this line up with what you would expect out of out of this bin? Yeah. I mean, my design brain is is, like, thinking of, like, a little bit of form layout, but I'm impressed that it's doing, like, half in full width, and it's kinda doing a really good first pass. Here's our awesome video. And, I don't have any. So let's see. That was my first concern is, like, you're not gonna have any, like, MP fours at the ready. We've got tons. Recdus MCP assets. What do we got? Blog post workflow. We could test the the TUS upload protocol at the same time. So, Meta, are you uploading an m c p p video for the m c p video? I was trying to. It looks like, some I'm getting some errors. So that's part of the fun of this. We just keep refreshing and and seeing more content or more data model show up. So we got post. Let's try this again. I'm gonna link this to my admin user. Here's our awesome video, And let's upload a video from device, see what it got. MCP assets. Do I have a video that is actually small enough? It looks like 30 megabytes is all I've got here. It feels like that that can work. Except I'm getting a issue. Oh. Oh. No stress. Alright. Let's just try an image for now. Oh, I've done something wrong with the images as well. I think TikTok has images, doesn't it? It it it does. It looks like something has gone wrong. I had reversed relationships. Okay. So now it's filling in the other relationships. And, unfortunately, I don't think it's gonna allow us to do this without a file, but we could solve for that really quickly. I know. I'm googling sample m p four file. I like cityscapes and all that on the mind I've seen. I think it is something that I've done here. Thirty seconds, 1.5 megs. 1.5 megs. Schema. It's like Add to proceed. Does this not know we are on a okay. So I spotted the issue, and that's a good point to bring up that this is really cool functionality to be able to prototype this quickly, but it's not a it's not a silver bullet. It looks like the MCP had tried to tag the wrong folder that did not exist inside the instance, and that's where the issue was coming from. So we could try this again. Here's our test post. Let's add admin user. Just going to I Slacked your URL if you wanna try doing it from URL. Okay. Alright. We'll try it from the URL. There we go. We've got the video file uploaded. We can set this to publish. We will save our post. And now that the MCP has done its thing, we got two minutes and forty seconds Infinite. To, let's just create create some test posts and sample content for each collection. Cool. Alright. In the meantime, Ben, what is great about Directus and how fast you could prototype like this? I mean, this is this is insane. Like, with all the things that we're encountering here I mean, even just the boilerplate stuff of getting the prompt right, trying to find just a sample video file. Like, I you just I feel like you can just, like, mess around. And the worst case scenario, you just command z it and then start again, or at least get, like, a good 80% of the way there. We got one minute and thirty six seconds. I think one of the the coolest things to showcase here is that we have a ready to go API. Right? And not just, I need, like, one API. We've got all the APIs for what we need on the back end of this thing. I love that it's doing all the colors and iconography too. Like, that to me is it just takes so much time. Like, oh, okay. What what, like, material symbol icon should I use for this or that? Like, that plus the colors and all the form layout. Like, that's Yeah. So let's go in and create a a like here. Here's our post. Cool. They created should be populated automatically. We got follows. We got comments. We can add a comment to the post. Test comment. Alright. User ID. Commenting on my own I got a user. Who said who said test comment? There we go. Alright. So thirty seconds left. I I feel like, you know, we didn't obviously do anything on the front end here, but now we we've got these APIs with a a little bit more time. We could probably slap together a front end, and have something up and running in under an hour, which, you know, the namesake of the show is is insane to me that you could do so much with DirectUs in an hour. But being able to get here in ten minutes, The entire back end is, like, the the schema. We've got, like, everything sort of configured the right way. We got the full API, REST, GraphQL. That's it. Ten minutes is up. Well, Ben, any kind of final words on Directus and AI before we we wrap this episode up? I mean, it's just it's super exciting. Like, it's the confluence of two big things. Like, AI is just, like, variable, ephemeral, like, crazy new emergent technology with all this potential, but it's kind of a little errant with all the craziness that can get thrown into the mix. And then you've got this, like, super stable, reliable, resilient back end platform of directors. I think, like, at the intersection of those two things is is, as Jonathan would say, there there's magic there. Yeah. No. This is this is awesome. And with that, we will end this episode. Thanks for joining us, and stay tuned for more. Thank you, Brian.","31c14d2e-45e1-4e6a-be3a-381f23d2362c",[221,222],"c1d5015e-188b-4fe4-9595-3f83f2e59241","c0d5cd0b-e341-4b10-8f91-d0efe6dc19a3",[],{"id":134,"number":135,"show":122,"year":136,"episodes":225},[138,139,140,141,142,143,144,145,146],{"id":143,"slug":227,"vimeo_id":228,"description":229,"tile":230,"length":231,"resources":8,"people":8,"episode_number":232,"published":154,"title":233,"video_transcript_html":234,"video_transcript_text":235,"content":8,"seo":236,"status":131,"episode_people":237,"recommendations":240,"season":241},"10-mins-menu-app","1138952700","Bryant and Rijk try to build the project of Rijk's dreams using the MCP, in 10 minutes or less!","9019e6ad-9dd3-42ab-9840-b39bf58fc5b7",16,6,"Building in 10 Minutes or Less: Menu App","\u003Cp>Speaker 0: Alright. Welcome back to an extra special episode of one app in ten minutes. I'm your host, Brian Gillespie, for Directus. Today, I honestly like, I I hope I'm not gonna get fired here, but, we are going to be building a mystery app with our CTO here at Directus, Reich Van Zanten. And I call this mystery app, and I say, hey.\u003C/p>\u003Cp>I hope I don't get fired in that, I've invited Reich to a call, but he has no idea he's gonna be on the show. And I call it mystery app because I have no idea what idea he's gonna throw out there or if it's even possible. Previously, the show was called 100 apps, hundred hours for this season or this spin off here. It is one app in ten minutes. The rules are we have ten minutes to plan and build an app.\u003C/p>\u003Cp>No more, no less. In this case, we're gonna be using what we do have at our disposal, which is Directus, claw.ai, and, Directus MCP connected. Welcome to one app ten minutes, Rick.\u003C/p>\u003Cp>Speaker 1: One app ten minutes.\u003C/p>\u003Cp>Speaker 0: For the audience, we are recording live. Did you have any idea that you're gonna be on this show today?\u003C/p>\u003Cp>Speaker 1: I had no clue. But here we\u003C/p>\u003Cp>Speaker 0: are. Amazing. Amazing. Alright. So I think you're already aware of 100 apps, hundred hours, one app, ten minutes, shortened time frame, leveraging AI.\u003C/p>\u003Cp>I've got mystery app here with Wrike. The world wants to know what what app are we gonna build today?\u003C/p>\u003Cp>Speaker 1: In ten minutes?\u003C/p>\u003Cp>Speaker 0: In ten minutes.\u003C/p>\u003Cp>Speaker 1: Oh, boy. Well, I could spend ten minutes thinking about an app, but that's not what I wanna do.\u003C/p>\u003Cp>Speaker 0: The most incredible cold opened ever, ladies and gentlemen. What a what a do you thought you were coming into, an afternoon meeting? Amazing.\u003C/p>\u003Cp>Speaker 1: Oh, one of many. One of many. Oh, this is I'm woefully ill prepared for this. Okay. We have ten minutes.\u003C/p>\u003Cp>Well, the time hasn't started yet,\u003C/p>\u003Cp>Speaker 0: has it? No. No. No. You got we gotta decide the app, then we got ten minutes to plan and build the thing, which is it is very ambitious in and of itself.\u003C/p>\u003Cp>Speaker 1: Yeah. I will say. Okay. Ten minutes. That makes it so much trickier.\u003C/p>\u003Cp>I was gonna say, I was recently working oh, maybe we can do a part of that. I was gonna say, I was recently working on sort of, like, a quizzing app.\u003C/p>\u003Cp>Speaker 0: A quizzing app. Okay.\u003C/p>\u003Cp>Speaker 1: Nights and weekends for, you know, getting, sending people a link. You could sign in. You can you can do a little quiz, and then we could check out the leaderboard. That is not a ten minute I don't ever at all. I'll tell you that.\u003C/p>\u003Cp>Oh. But maybe it is. Let's see here. Okay.\u003C/p>\u003Cp>Speaker 0: No pressure.\u003C/p>\u003Cp>Speaker 1: None whatsoever.\u003C/p>\u003Cp>Speaker 0: To give you context, we've already done a lot of the former 100 apps, hundred hours. So Expensify clone, Airbnb, Netflix. Like, already done all these things. Multi site CMS. All the all the expected fruit is is already picked off the tree.\u003C/p>\u003Cp>Speaker 1: You're not making it any easier. Wait. If this is a if if we're gonna be using AI to build this stuff, why don't we use AI to come up with what we're building in the first place?\u003C/p>\u003Cp>Speaker 0: Alright. Alright. Alright. I've got\u003C/p>\u003Cp>Speaker 1: I've got ten minutes.\u003C/p>\u003Cp>Speaker 0: Rike Van Zanten, our CTO, on an episode of one app, ten minutes. What should we build? Bum bum bum. I'm Curious to see if it calls the actual direct SMTP here, but I I don't think so. Real time dashboard builder.\u003C/p>\u003Cp>Oh, it it knows you. You're a known quantity. Clearly. Schema visualizer, webhook, ecommerce, headless ecommerce, QR code menu generator, team mood tracker, meeting room optimizer.\u003C/p>\u003Cp>Speaker 1: I think the QR menu one is actually not too bad of an idea. Right?\u003C/p>\u003Cp>Speaker 0: Because this is a restaurant.\u003C/p>\u003Cp>Speaker 1: Thing or you're going in, you're sitting down at the table, you have to scan a little QR code, you pull up a menu on your phone. So very remote. Where does this menu come from? Where does it go? Could be something.\u003C/p>\u003Cp>Speaker 0: Alright. Restaurant menu. QR code is doing everything.\u003C/p>\u003Cp>Speaker 1: Ten minutes where I create a new startup on the fly.\u003C/p>\u003Cp>Speaker 0: Restaurant menu, QR code, generator thingy. There we go. Alright. We're ready? We're starting the clock.\u003C/p>\u003Cp>What do we need out of this thing? Obviously Well,\u003C/p>\u003Cp>Speaker 1: we need menus. We need things on the menu. So we need menu items. We need ideally some sort of, like, allergy information for every menu item. So that is, like, ingredients maybe, or, like, allergy information, or components.\u003C/p>\u003Cp>Like, sometimes on menus, you see, you know, the menu item and then individual parts of what's in it. Right? Yep. So yesterday night, at least four times of what is in the Negroni. Just kidding.\u003C/p>\u003Cp>What else do we need? We need well, we're gonna be generating a QR code for every menu, I guess, because we're rendering out each menu on a separate page. So that makes sense.\u003C/p>\u003Cp>Speaker 0: That QR code?\u003C/p>\u003Cp>Speaker 1: Oh, we need a a sort of category for the menu item. Like, is it an entree, or is it an appetizer, or is it a dessert? And then on the menu itself, we probably need a similar thing for, is it a lunch menu, or a dinner menu, or some special event? You know, this is like a Christmas day, menu or something.\u003C/p>\u003Cp>Speaker 0: So what would that be?\u003C/p>\u003Cp>Speaker 1: I was already thinking Yeah. Type a menu. Yeah. Holiday special.\u003C/p>\u003Cp>Speaker 0: Holiday. Alright. Now, I've already set the audience up for this, but what we've got already is I've got a blank direct to sentence. We've got the Claude AI connected via the direct to MCP, and let's build the menu system. So I'm just gonna dump this in here.\u003C/p>\u003Cp>Again, we're on the clock. So, you know, we're just gonna oh, gosh. Okay. That's\u003C/p>\u003Cp>Speaker 1: Oh, no.\u003C/p>\u003Cp>Speaker 0: That's not even formatted properly. Okay. Allergy category for the menu. Flow for during QR code. Alright.\u003C/p>\u003Cp>This is the magic moment. Can you one shot a brand new startup in seven minutes and fifty five seconds? We'll see.\u003C/p>\u003Cp>Speaker 1: Well, in no context either. I mean, you're just you're giving it nothing. You're truly giving it nothing.\u003C/p>\u003Cp>Speaker 0: Absolutely nothing. Right? So the Directed MCP, we've got a bunch of tools that are preprogrammed into this thing. And you'll see, like, the system prompt here, which is basically just some information about, how the assistant should act. Then we have a tool that allows it to introspect the direct to schema or or not even introspect, just see the direct to schema.\u003C/p>\u003Cp>And here it's already gone to work. So I'm just gonna hit refresh over here on the left. And by the power of AI combined, we now have some stuff. So we can start creating a new menu. This is, Reich's special lunch menu.\u003C/p>\u003Cp>And is this, is this something that we would expect to see? I mean, how would you grade this? So we got the type of lunch. That's pretty solid. Is this active?\u003C/p>\u003Cp>There's the QR code for the menu. Uh-oh. Doesn't like that, does it?\u003C/p>\u003Cp>Speaker 1: Is it still create? Yeah.\u003C/p>\u003Cp>Speaker 0: It is still adding our relationships. So maybe that's where\u003C/p>\u003Cp>Speaker 1: I would I would do it a second until it's done, frankly.\u003C/p>\u003Cp>Speaker 0: Trying to get ahead of ourselves.\u003C/p>\u003Cp>Speaker 1: Because we only have so we only have so many seconds. But\u003C/p>\u003Cp>Speaker 0: There we go. Alright. So one thing that that I really like about this is just seeing, like, the volume of work that I would have to to do myself. Now it like, it's not that much work to spin up a back end with Directus. And the the visual, like, drag and drop the data model part of it is beautiful.\u003C/p>\u003Cp>But when you're watching a machine do all of this work for you, man, that's it's even better.\u003C/p>\u003Cp>Speaker 1: And, man, what can I say? What what is better than doing work is just not doing work.\u003C/p>\u003Cp>Speaker 0: Not doing any work. Alright. So we're gonna call this, like, special lunch menu. Now let's see if this actually works this time around. There we go.\u003C/p>\u003Cp>Alright. So we've got a menu. Looks like it's already added some menu items. How do you feel\u003C/p>\u003Cp>Speaker 1: about be seeing eating jumbo shrimp for lunch or Atlantic salmon for that, man.\u003C/p>\u003Cp>Speaker 0: Alright. We've got our ingredients. Looks like we've got an allergens. Okay. We got a preset there.\u003C/p>\u003Cp>And then we got our categories for appetizers, mains, desserts, beverages. What is this thing doing? Okay. It's\u003C/p>\u003Cp>Speaker 1: looking like I think it's building the front end for it now too.\u003C/p>\u003Cp>Speaker 0: Wow. It's actually building the front end. I didn't even ask that. But let's just let's check on the the flow progress. Right?\u003C/p>\u003Cp>We got five minutes left, which is, actually, I'm I'm surprised we're even here at this point.\u003C/p>\u003Cp>Speaker 1: But I'll say stop the clock because I'm seeing a menu on the screen right now.\u003C/p>\u003Cp>Speaker 0: Right. Is it is it actually calling Directus too? Oh, no. Of course. Yeah.\u003C/p>\u003Cp>Yeah.\u003C/p>\u003Cp>Speaker 1: Yeah. Okay. It's just gonna send that be too good to be true. Yeah.\u003C/p>\u003Cp>Speaker 0: Yeah.\u003C/p>\u003Cp>Speaker 1: Alright. So let's talk easy enough to do, though. This is one SDK call and you're in in business.\u003C/p>\u003Cp>Speaker 0: Yeah. So let's toggle this and see. Generate QR code for the menu. Okay. So we've generated the QR code.\u003C/p>\u003Cp>Looks like it might have missed a few steps. We generated a QR code. Is this even valid? What is the QR server?\u003C/p>\u003Cp>Speaker 1: I don't think I've ever seen this. Is that a thing? QR is a dummy. I think that's just a dummy, isn't it?\u003C/p>\u003Cp>Speaker 0: Yeah. Yeah. QR code, API. Can we get something really quickly? Yeah.\u003C/p>\u003Cp>Let's see here\u003C/p>\u003Cp>Speaker 1: another thing. We can oh, yeah. Maybe.\u003C/p>\u003Cp>Speaker 0: We're just gonna straight up vibe code. Gotta go back. Gotta go back. Gotta go back. Fix the flow to import the generated.\u003C/p>\u003Cp>Speaker 1: Oh, yeah. Stop making typos, Brian. We're on the clock. Come on.\u003C/p>\u003Cp>Speaker 0: The QR code generated QR code. Here's the service. What's the endpoint? What's the endpoint?\u003C/p>\u003Cp>Speaker 1: I can find it. I'm sure I can do it.\u003C/p>\u003Cp>Speaker 0: This is so much more fun with a partner.\u003C/p>\u003Cp>Speaker 1: Brian, let's go. Let's go. Let's go. Let's go. Come on.\u003C/p>\u003Cp>Come on.\u003C/p>\u003Cp>Speaker 0: Files. Come on. Import. Files import.\u003C/p>\u003Cp>Speaker 1: If we do as an import file operation, is that a thing? Is that even exist?\u003C/p>\u003Cp>Speaker 0: We do not have that.\u003C/p>\u003Cp>Speaker 1: We should.\u003C/p>\u003Cp>Speaker 0: Can you\u003C/p>\u003Cp>Speaker 1: build that in ten minutes? Same. We could.\u003C/p>\u003Cp>Speaker 0: Import file URL. Hit point. Oh, it's still building. Okay. It's reconnecting.\u003C/p>\u003Cp>Reconnecting. Alright. So it's doing some some type of work on our behalf here. QR code. Oh, no.\u003C/p>\u003Cp>Dun dun dun. Two minutes still. Alright. What do we what do we got here? Okay.\u003C/p>\u003Cp>Still not generating the QR code. Okay. Well, I think one thing we can agree on here\u003C/p>\u003Cp>Speaker 1: I mean, a QR code, you can easily generate on the client side. There's a bunch of libraries for that kind of stuff. So that that's the least of my concerns, frankly.\u003C/p>\u003Cp>Speaker 0: Least of the concerns. Alright. What other concerns do we have here?\u003C/p>\u003Cp>Speaker 1: Well, let's just quickly check on the data model. Right? Because we have we have menus, we have items, we have ingredients, we have the allergens on the ingredients. So all this is coming together really, really nicely.\u003C/p>\u003Cp>Speaker 0: Shall we, like, look at the API calls? Because, obviously, you're gonna render this on the front end.\u003C/p>\u003Cp>Speaker 1: Yeah. Did you give it a, like, public access or is it already?\u003C/p>\u003Cp>Speaker 0: I've just logged in. So we can make this let's let's yeah. Maybe showcase that really quickly. Alright. I can make all this publicly accessible just by going to our access policies and adding all these collections.\u003C/p>\u003Cp>I need faster fingers. There we go. Alright. We go back. Items, menus, fields.\u003C/p>\u003Cp>There we go. We've got our lunch menu. Oh, I forgot to add the the items to the menu. Can we do this? Where's the menu?\u003C/p>\u003Cp>Menu. Right? Special lunch menu. Add our existing items. Where we at?\u003C/p>\u003Cp>Twenty five seconds.\u003C/p>\u003Cp>Speaker 1: Hit save.\u003C/p>\u003Cp>Speaker 0: Items, menus, fields. Okay. I don't see the menu items. We need menu items dot asterisk.\u003C/p>\u003Cp>Speaker 1: There we go.\u003C/p>\u003Cp>Speaker 0: Junction table.\u003C/p>\u003Cp>Speaker 1: Items. Double star it. Double star.\u003C/p>\u003Cp>Speaker 0: Double star. No. Well, that's it. That's the ten minutes. We have built a menu generating app thingy.\u003C/p>\u003Cp>Ten minutes or less. How do we feel about this?\u003C/p>\u003Cp>Speaker 1: It's pretty cool. I will say. My favorite is the icons and the colors for the for the collection.\u003C/p>\u003Cp>Speaker 0: I so that again, like, this is all AI, you know, there's a lot of hype out there. But one thing that that I appreciate about the MCP that we built here is there's real value to this, in my opinion, of how quick can you prototype, how fast can you iterate on this. In ten minutes or less, completely fresh. No idea. There it is.\u003C/p>\u003Cp>Restaurant menu app down. Reich, anything, anything you'd like to add as we we close this episode, this chaotic episode, if you didn't need enough drama and one app in ten minutes bringing in somebody totally unexpectedly. Well, Reich, thank you for joining me for this. You know, again, I think this is a lot of fun, but again, it just goes to show how far and how fast you could get when you combine Directus and AI. That's it, folks.\u003C/p>\u003Cp>Make sure you stay tuned for more episodes of one app, ten minutes.\u003C/p>","Alright. Welcome back to an extra special episode of one app in ten minutes. I'm your host, Brian Gillespie, for Directus. Today, I honestly like, I I hope I'm not gonna get fired here, but, we are going to be building a mystery app with our CTO here at Directus, Reich Van Zanten. And I call this mystery app, and I say, hey. I hope I don't get fired in that, I've invited Reich to a call, but he has no idea he's gonna be on the show. And I call it mystery app because I have no idea what idea he's gonna throw out there or if it's even possible. Previously, the show was called 100 apps, hundred hours for this season or this spin off here. It is one app in ten minutes. The rules are we have ten minutes to plan and build an app. No more, no less. In this case, we're gonna be using what we do have at our disposal, which is Directus, claw.ai, and, Directus MCP connected. Welcome to one app ten minutes, Rick. One app ten minutes. For the audience, we are recording live. Did you have any idea that you're gonna be on this show today? I had no clue. But here we are. Amazing. Amazing. Alright. So I think you're already aware of 100 apps, hundred hours, one app, ten minutes, shortened time frame, leveraging AI. I've got mystery app here with Wrike. The world wants to know what what app are we gonna build today? In ten minutes? In ten minutes. Oh, boy. Well, I could spend ten minutes thinking about an app, but that's not what I wanna do. The most incredible cold opened ever, ladies and gentlemen. What a what a do you thought you were coming into, an afternoon meeting? Amazing. Oh, one of many. One of many. Oh, this is I'm woefully ill prepared for this. Okay. We have ten minutes. Well, the time hasn't started yet, has it? No. No. No. You got we gotta decide the app, then we got ten minutes to plan and build the thing, which is it is very ambitious in and of itself. Yeah. I will say. Okay. Ten minutes. That makes it so much trickier. I was gonna say, I was recently working oh, maybe we can do a part of that. I was gonna say, I was recently working on sort of, like, a quizzing app. A quizzing app. Okay. Nights and weekends for, you know, getting, sending people a link. You could sign in. You can you can do a little quiz, and then we could check out the leaderboard. That is not a ten minute I don't ever at all. I'll tell you that. Oh. But maybe it is. Let's see here. Okay. No pressure. None whatsoever. To give you context, we've already done a lot of the former 100 apps, hundred hours. So Expensify clone, Airbnb, Netflix. Like, already done all these things. Multi site CMS. All the all the expected fruit is is already picked off the tree. You're not making it any easier. Wait. If this is a if if we're gonna be using AI to build this stuff, why don't we use AI to come up with what we're building in the first place? Alright. Alright. Alright. I've got I've got ten minutes. Rike Van Zanten, our CTO, on an episode of one app, ten minutes. What should we build? Bum bum bum. I'm Curious to see if it calls the actual direct SMTP here, but I I don't think so. Real time dashboard builder. Oh, it it knows you. You're a known quantity. Clearly. Schema visualizer, webhook, ecommerce, headless ecommerce, QR code menu generator, team mood tracker, meeting room optimizer. I think the QR menu one is actually not too bad of an idea. Right? Because this is a restaurant. Thing or you're going in, you're sitting down at the table, you have to scan a little QR code, you pull up a menu on your phone. So very remote. Where does this menu come from? Where does it go? Could be something. Alright. Restaurant menu. QR code is doing everything. Ten minutes where I create a new startup on the fly. Restaurant menu, QR code, generator thingy. There we go. Alright. We're ready? We're starting the clock. What do we need out of this thing? Obviously Well, we need menus. We need things on the menu. So we need menu items. We need ideally some sort of, like, allergy information for every menu item. So that is, like, ingredients maybe, or, like, allergy information, or components. Like, sometimes on menus, you see, you know, the menu item and then individual parts of what's in it. Right? Yep. So yesterday night, at least four times of what is in the Negroni. Just kidding. What else do we need? We need well, we're gonna be generating a QR code for every menu, I guess, because we're rendering out each menu on a separate page. So that makes sense. That QR code? Oh, we need a a sort of category for the menu item. Like, is it an entree, or is it an appetizer, or is it a dessert? And then on the menu itself, we probably need a similar thing for, is it a lunch menu, or a dinner menu, or some special event? You know, this is like a Christmas day, menu or something. So what would that be? I was already thinking Yeah. Type a menu. Yeah. Holiday special. Holiday. Alright. Now, I've already set the audience up for this, but what we've got already is I've got a blank direct to sentence. We've got the Claude AI connected via the direct to MCP, and let's build the menu system. So I'm just gonna dump this in here. Again, we're on the clock. So, you know, we're just gonna oh, gosh. Okay. That's Oh, no. That's not even formatted properly. Okay. Allergy category for the menu. Flow for during QR code. Alright. This is the magic moment. Can you one shot a brand new startup in seven minutes and fifty five seconds? We'll see. Well, in no context either. I mean, you're just you're giving it nothing. You're truly giving it nothing. Absolutely nothing. Right? So the Directed MCP, we've got a bunch of tools that are preprogrammed into this thing. And you'll see, like, the system prompt here, which is basically just some information about, how the assistant should act. Then we have a tool that allows it to introspect the direct to schema or or not even introspect, just see the direct to schema. And here it's already gone to work. So I'm just gonna hit refresh over here on the left. And by the power of AI combined, we now have some stuff. So we can start creating a new menu. This is, Reich's special lunch menu. And is this, is this something that we would expect to see? I mean, how would you grade this? So we got the type of lunch. That's pretty solid. Is this active? There's the QR code for the menu. Uh-oh. Doesn't like that, does it? Is it still create? Yeah. It is still adding our relationships. So maybe that's where I would I would do it a second until it's done, frankly. Trying to get ahead of ourselves. Because we only have so we only have so many seconds. But There we go. Alright. So one thing that that I really like about this is just seeing, like, the volume of work that I would have to to do myself. Now it like, it's not that much work to spin up a back end with Directus. And the the visual, like, drag and drop the data model part of it is beautiful. But when you're watching a machine do all of this work for you, man, that's it's even better. And, man, what can I say? What what is better than doing work is just not doing work. Not doing any work. Alright. So we're gonna call this, like, special lunch menu. Now let's see if this actually works this time around. There we go. Alright. So we've got a menu. Looks like it's already added some menu items. How do you feel about be seeing eating jumbo shrimp for lunch or Atlantic salmon for that, man. Alright. We've got our ingredients. Looks like we've got an allergens. Okay. We got a preset there. And then we got our categories for appetizers, mains, desserts, beverages. What is this thing doing? Okay. It's looking like I think it's building the front end for it now too. Wow. It's actually building the front end. I didn't even ask that. But let's just let's check on the the flow progress. Right? We got five minutes left, which is, actually, I'm I'm surprised we're even here at this point. But I'll say stop the clock because I'm seeing a menu on the screen right now. Right. Is it is it actually calling Directus too? Oh, no. Of course. Yeah. Yeah. Yeah. Okay. It's just gonna send that be too good to be true. Yeah. Yeah. Alright. So let's talk easy enough to do, though. This is one SDK call and you're in in business. Yeah. So let's toggle this and see. Generate QR code for the menu. Okay. So we've generated the QR code. Looks like it might have missed a few steps. We generated a QR code. Is this even valid? What is the QR server? I don't think I've ever seen this. Is that a thing? QR is a dummy. I think that's just a dummy, isn't it? Yeah. Yeah. QR code, API. Can we get something really quickly? Yeah. Let's see here another thing. We can oh, yeah. Maybe. We're just gonna straight up vibe code. Gotta go back. Gotta go back. Gotta go back. Fix the flow to import the generated. Oh, yeah. Stop making typos, Brian. We're on the clock. Come on. The QR code generated QR code. Here's the service. What's the endpoint? What's the endpoint? I can find it. I'm sure I can do it. This is so much more fun with a partner. Brian, let's go. Let's go. Let's go. Let's go. Come on. Come on. Files. Come on. Import. Files import. If we do as an import file operation, is that a thing? Is that even exist? We do not have that. We should. Can you build that in ten minutes? Same. We could. Import file URL. Hit point. Oh, it's still building. Okay. It's reconnecting. Reconnecting. Alright. So it's doing some some type of work on our behalf here. QR code. Oh, no. Dun dun dun. Two minutes still. Alright. What do we what do we got here? Okay. Still not generating the QR code. Okay. Well, I think one thing we can agree on here I mean, a QR code, you can easily generate on the client side. There's a bunch of libraries for that kind of stuff. So that that's the least of my concerns, frankly. Least of the concerns. Alright. What other concerns do we have here? Well, let's just quickly check on the data model. Right? Because we have we have menus, we have items, we have ingredients, we have the allergens on the ingredients. So all this is coming together really, really nicely. Shall we, like, look at the API calls? Because, obviously, you're gonna render this on the front end. Yeah. Did you give it a, like, public access or is it already? I've just logged in. So we can make this let's let's yeah. Maybe showcase that really quickly. Alright. I can make all this publicly accessible just by going to our access policies and adding all these collections. I need faster fingers. There we go. Alright. We go back. Items, menus, fields. There we go. We've got our lunch menu. Oh, I forgot to add the the items to the menu. Can we do this? Where's the menu? Menu. Right? Special lunch menu. Add our existing items. Where we at? Twenty five seconds. Hit save. Items, menus, fields. Okay. I don't see the menu items. We need menu items dot asterisk. There we go. Junction table. Items. Double star it. Double star. Double star. No. Well, that's it. That's the ten minutes. We have built a menu generating app thingy. Ten minutes or less. How do we feel about this? It's pretty cool. I will say. My favorite is the icons and the colors for the for the collection. I so that again, like, this is all AI, you know, there's a lot of hype out there. But one thing that that I appreciate about the MCP that we built here is there's real value to this, in my opinion, of how quick can you prototype, how fast can you iterate on this. In ten minutes or less, completely fresh. No idea. There it is. Restaurant menu app down. Reich, anything, anything you'd like to add as we we close this episode, this chaotic episode, if you didn't need enough drama and one app in ten minutes bringing in somebody totally unexpectedly. Well, Reich, thank you for joining me for this. You know, again, I think this is a lot of fun, but again, it just goes to show how far and how fast you could get when you combine Directus and AI. That's it, folks. Make sure you stay tuned for more episodes of one app, ten minutes.","7ea951b9-ee5d-456e-8e38-ea63fc3010f0",[238,239],"6dcc99c9-1ea6-46e3-a22e-8b730be473a3","8e8ddf61-cf6f-4a94-99fd-57c07f3b26b8",[],{"id":134,"number":135,"show":122,"year":136,"episodes":242},[138,139,140,141,142,143,144,145,146],{"id":144,"slug":244,"vimeo_id":245,"description":246,"tile":247,"length":248,"resources":8,"people":8,"episode_number":249,"published":154,"title":250,"video_transcript_html":251,"video_transcript_text":252,"content":8,"seo":253,"status":131,"episode_people":254,"recommendations":255,"season":256},"mcp-build-note-app","1138970580","Joshua Bemendorfer talks through his project: building a note-taking app.","51d2f9d1-b2b2-420f-bf86-e7c7dba6926e",9,7,"Community MCP Build: Note-Taking App","\u003Cp>Speaker 0: Today, I would like to show you how to supercharge AI batch processing using Directus MCP with a self learning note taking workflow. Batch processing is really effective for, for situations where you need raw speed, but it kinda breaks down when you need to make creative decisions on individual items and on the batch as a whole. You can't really do that without a human element. AI models, I found, are great for those sorts of tasks, but they break down when you give them a whole bunch of data. And they also struggle with finding the right information to pull in to make those creative decisions.\u003C/p>\u003Cp>So the solution I found is to teach AI to take notes. By implementing a workflow that lets Directus MCP take notes directly on whatever it's working on, we gain the ability to do things like long running task tracking because the model can pick up where it left off and and, start fresh with a new context window. And we also get things like self improving database access. You take a note on all the different things that you ran into difficulty with and what you did to solve them and how it worked out. And it speeds up future runs because it can read that note and go, okay.\u003C/p>\u003Cp>So don't do this. Do this, and I'll be able to just continue with the task I was working on. And then by letting it take notes, we can consolidate a whole bunch of data that was stored in the database. Say, Say, for example, you wanted to pull out all of the aliases for your articles that are related to rabbits and you also want the ID for each one. Well, you can just dump that in a note and analyze it in future runs.\u003C/p>\u003Cp>So in our case, Directus is the knowledge backbone. Every note is just a key value pair. I'll go ahead and show you the notes table here. It's a very simple setup. We have a key, which is just a string, and this is what the AI model uses to kind of categorize what the different notes are about.\u003C/p>\u003Cp>And we have the value, and that's just a big markdown, field. It's a very simple setup, but it's surprisingly powerful. By using semantic keys and explicit instructions on how to record and review notes, we give the AI models a very quick and effective way to find the context they need. So this database has a problem. We have a whole bunch of articles.\u003C/p>\u003Cp>We have a 124 articles. The titles don't really make any sense. Oh, the bodies are in Latin. And I I can't work with this. So I've created a prompt for Claude to read the article titles and come up with a concept proposal for each one.\u003C/p>\u003Cp>What could this article actually be about? So let's look at the system prompt here. You can see that there's two things that it's informing the AI model about. That there's an AI notes table in the rough structure of that table and that it needs to read and record its database insights, especially how it solved problems it ran into and just update that note on every run. This gives it the self learning ability we were talking about earlier.\u003C/p>\u003Cp>It's able to figure out what it ran into last time and fix it on the next run. Now back to the task. I've told the AI model it's a skilled content analyst and writer. We have a huge collection of articles. All the bodies are full of nonsense, all that.\u003C/p>\u003Cp>Alright. So its job is to record the concepts that it comes up with in an article concepts note. It's not sure what to do with an article based on the title. Be creative. Find a wacky concept.\u003C/p>\u003Cp>Then we have some very specific instructions. We say always read the template and template article concepts, this is a note, to determine how to structure your output. Always create a new article concepts note for each run. Always give it the exact name. Always process 30 articles.\u003C/p>\u003Cp>Always review the last note you created before starting to make sure you don't duplicate any work. This template lets us make sure that we get consistent output on every run. Alright. Let's go ahead and run this a few times and see what its output looks like. I turn on auto refresh, and then I'm gonna go into Claude, add a prompt from Directus, and examine articles and propose concepts.\u003C/p>\u003Cp>If you want to know how to use this feature, go ahead and take a look at the Directus MCP documentation, and then we're just gonna send it. Now one of the fun things you can do with read and write access direct to Directus is if the model makes a mistake, you can ask it to update the prompt to fix its mistake in future runs. Article concepts. It has identified directly that just about every article appears to be lorem ipsum gibberish and occasional test content. Alright.\u003C/p>\u003Cp>So came up with a bunch of different articles on a bunch of different ideas. In a bit, we're gonna use this to actually write all these different article concepts. But for now, I just wanna keep going through the batch process to show you how it's able to pick up where it left off and continue. On this run, if you notice, it actually picked up that it needed to look up the translations and didn't have to figure that out from the schema this time. Alright.\u003C/p>\u003Cp>It is a new day. My cloud usage limits have reset, and we are ready to continue with step two, which is generating the articles based on the concepts that we've put together. All of the, article concepts have now been saved in these, concept notes. They take the article title. They try to figure out what on earth, the article should be about and propose a concept.\u003C/p>\u003Cp>So now we have the next step of the process, which is to generate a actual article for each of the articles we generate concepts on in both English and German. And we're gonna go in batches of 10 and see how well that works. We'll take this prompt and it should automatically read over all of our article concept notes and start filling in articles for those. Let's go. So we're going to use the turn concepts into full fledged articles prompt.\u003C/p>\u003Cp>That's going to go and read through all of our article concept notes and start filling in those articles, and I'll show you those articles as it writes them. I have no idea if these articles are going to make any sense. This is going to be fun. I'm having fun reading through these articles. I have no idea how helpful any of this is, but it at least sounds very convincing.\u003C/p>\u003Cp>Now, the key advantage to this approach is that if we were generating concepts and writing articles at the same time, we'd have to use much smaller batches. But because we split the process into two steps, we've saved a ton of context window, and we're able to work in larger batches. And as a bonus, we can do multiple things with those article summaries, while we're working on generating those articles from those summaries. Now, this isn't the most efficient workflow for generating articles. There's way better workflows for that.\u003C/p>\u003Cp>We're just demonstrating this concept. But a nice thing about this approach is that you can perform multiple tasks at the same time. For example, while we were working on these articles, we could also be working on something completely different using the notes that we're generating the articles from. I have this prompt here, categorize article concepts, which will allow Claude to suggest an article taxonomy based on the summaries it's already generated. It'll just read all the article concept notes and create a new note suggesting that new taxonomy.\u003C/p>\u003Cp>I like the way it describes these articles as a fascinating collection of technical and business focused article concepts with creative jargon filled titles that have been transformed into practical valuable content ideas. I'm not quite sure I'd be so positive about it, but it gets the idea across. Alright. Let's take a look at the categories that it generated and the opportunities that identified. So here's some content gaps, AI and ethics, sustainable technology operations, human centered digital transformation, all sorts of other different things.\u003C/p>\u003Cp>And here are some of the categories that's come up with. System architecture, it's put a bunch of different articles in that category, Ecommerce, digital business, data management and analytics, user experience and interface design, business operations or business optimization management, automation and AI systems, project management and collaboration. So it's basically taken all of those articles that we put together and categorized them according to common themes in those articles. This could be really useful if you have an existing content collection that you're trying to build a new categorization system for. And what's great is you could just take this and create another prompt to actually apply that taxonomy to articles, and create categories and all sorts of different things for that in Directus.\u003C/p>\u003Cp>So by the end of this run, here's what the AI model has built: concept proposals for every article in our database, full body content for many of those articles in English and German, a whole new content categorization system for the articles it wrote, progress logs so that it doesn't lose its place, and database access hints that documents how the challenges were solved so that on future runs, it could do even better. Instead of treating AI like a disposable worker or one that can infinitely stuff and stuff and stuff and work on something until it gets all confused, we're treating it like a teammate that tracks and follows up on what it's done across multiple days, multiple runs, starting with a clean slate on each task and only pulling in the information that it needs. We turned a human in the loop batch processing task into an AI creative workflow by splitting it into discrete steps and letting it kind of pick up where it left off. We gave the model a way to track progress. We built templates to make the results more consistent and reusable, and we recorded information about the database to help the AI model get smarter over time.\u003C/p>\u003Cp>I've had a ton of fun nailing down this workflow and testing it on all sorts of tasks. For my work, we're actually using it to do things similar to this where we're analyzing a huge amount of content and trying to figure out how to organize and categorize and and do all sorts of things, and it's it's doing really well. The ability to take notes, record what it's done, and just kind of start over and create its own context has been incredibly helpful with turning a workflow that works really well for one or two articles into something that works well across thousands. So thank you for your time. I really look forward to seeing what everyone does with Directus and Directus MCP going forward, especially as Directus MCP evolves, AI models get more capable, and just the overall core gets stronger.\u003C/p>\u003Cp>The future is gonna be fun.\u003C/p>","Today, I would like to show you how to supercharge AI batch processing using Directus MCP with a self learning note taking workflow. Batch processing is really effective for, for situations where you need raw speed, but it kinda breaks down when you need to make creative decisions on individual items and on the batch as a whole. You can't really do that without a human element. AI models, I found, are great for those sorts of tasks, but they break down when you give them a whole bunch of data. And they also struggle with finding the right information to pull in to make those creative decisions. So the solution I found is to teach AI to take notes. By implementing a workflow that lets Directus MCP take notes directly on whatever it's working on, we gain the ability to do things like long running task tracking because the model can pick up where it left off and and, start fresh with a new context window. And we also get things like self improving database access. You take a note on all the different things that you ran into difficulty with and what you did to solve them and how it worked out. And it speeds up future runs because it can read that note and go, okay. So don't do this. Do this, and I'll be able to just continue with the task I was working on. And then by letting it take notes, we can consolidate a whole bunch of data that was stored in the database. Say, Say, for example, you wanted to pull out all of the aliases for your articles that are related to rabbits and you also want the ID for each one. Well, you can just dump that in a note and analyze it in future runs. So in our case, Directus is the knowledge backbone. Every note is just a key value pair. I'll go ahead and show you the notes table here. It's a very simple setup. We have a key, which is just a string, and this is what the AI model uses to kind of categorize what the different notes are about. And we have the value, and that's just a big markdown, field. It's a very simple setup, but it's surprisingly powerful. By using semantic keys and explicit instructions on how to record and review notes, we give the AI models a very quick and effective way to find the context they need. So this database has a problem. We have a whole bunch of articles. We have a 124 articles. The titles don't really make any sense. Oh, the bodies are in Latin. And I I can't work with this. So I've created a prompt for Claude to read the article titles and come up with a concept proposal for each one. What could this article actually be about? So let's look at the system prompt here. You can see that there's two things that it's informing the AI model about. That there's an AI notes table in the rough structure of that table and that it needs to read and record its database insights, especially how it solved problems it ran into and just update that note on every run. This gives it the self learning ability we were talking about earlier. It's able to figure out what it ran into last time and fix it on the next run. Now back to the task. I've told the AI model it's a skilled content analyst and writer. We have a huge collection of articles. All the bodies are full of nonsense, all that. Alright. So its job is to record the concepts that it comes up with in an article concepts note. It's not sure what to do with an article based on the title. Be creative. Find a wacky concept. Then we have some very specific instructions. We say always read the template and template article concepts, this is a note, to determine how to structure your output. Always create a new article concepts note for each run. Always give it the exact name. Always process 30 articles. Always review the last note you created before starting to make sure you don't duplicate any work. This template lets us make sure that we get consistent output on every run. Alright. Let's go ahead and run this a few times and see what its output looks like. I turn on auto refresh, and then I'm gonna go into Claude, add a prompt from Directus, and examine articles and propose concepts. If you want to know how to use this feature, go ahead and take a look at the Directus MCP documentation, and then we're just gonna send it. Now one of the fun things you can do with read and write access direct to Directus is if the model makes a mistake, you can ask it to update the prompt to fix its mistake in future runs. Article concepts. It has identified directly that just about every article appears to be lorem ipsum gibberish and occasional test content. Alright. So came up with a bunch of different articles on a bunch of different ideas. In a bit, we're gonna use this to actually write all these different article concepts. But for now, I just wanna keep going through the batch process to show you how it's able to pick up where it left off and continue. On this run, if you notice, it actually picked up that it needed to look up the translations and didn't have to figure that out from the schema this time. Alright. It is a new day. My cloud usage limits have reset, and we are ready to continue with step two, which is generating the articles based on the concepts that we've put together. All of the, article concepts have now been saved in these, concept notes. They take the article title. They try to figure out what on earth, the article should be about and propose a concept. So now we have the next step of the process, which is to generate a actual article for each of the articles we generate concepts on in both English and German. And we're gonna go in batches of 10 and see how well that works. We'll take this prompt and it should automatically read over all of our article concept notes and start filling in articles for those. Let's go. So we're going to use the turn concepts into full fledged articles prompt. That's going to go and read through all of our article concept notes and start filling in those articles, and I'll show you those articles as it writes them. I have no idea if these articles are going to make any sense. This is going to be fun. I'm having fun reading through these articles. I have no idea how helpful any of this is, but it at least sounds very convincing. Now, the key advantage to this approach is that if we were generating concepts and writing articles at the same time, we'd have to use much smaller batches. But because we split the process into two steps, we've saved a ton of context window, and we're able to work in larger batches. And as a bonus, we can do multiple things with those article summaries, while we're working on generating those articles from those summaries. Now, this isn't the most efficient workflow for generating articles. There's way better workflows for that. We're just demonstrating this concept. But a nice thing about this approach is that you can perform multiple tasks at the same time. For example, while we were working on these articles, we could also be working on something completely different using the notes that we're generating the articles from. I have this prompt here, categorize article concepts, which will allow Claude to suggest an article taxonomy based on the summaries it's already generated. It'll just read all the article concept notes and create a new note suggesting that new taxonomy. I like the way it describes these articles as a fascinating collection of technical and business focused article concepts with creative jargon filled titles that have been transformed into practical valuable content ideas. I'm not quite sure I'd be so positive about it, but it gets the idea across. Alright. Let's take a look at the categories that it generated and the opportunities that identified. So here's some content gaps, AI and ethics, sustainable technology operations, human centered digital transformation, all sorts of other different things. And here are some of the categories that's come up with. System architecture, it's put a bunch of different articles in that category, Ecommerce, digital business, data management and analytics, user experience and interface design, business operations or business optimization management, automation and AI systems, project management and collaboration. So it's basically taken all of those articles that we put together and categorized them according to common themes in those articles. This could be really useful if you have an existing content collection that you're trying to build a new categorization system for. And what's great is you could just take this and create another prompt to actually apply that taxonomy to articles, and create categories and all sorts of different things for that in Directus. So by the end of this run, here's what the AI model has built: concept proposals for every article in our database, full body content for many of those articles in English and German, a whole new content categorization system for the articles it wrote, progress logs so that it doesn't lose its place, and database access hints that documents how the challenges were solved so that on future runs, it could do even better. Instead of treating AI like a disposable worker or one that can infinitely stuff and stuff and stuff and work on something until it gets all confused, we're treating it like a teammate that tracks and follows up on what it's done across multiple days, multiple runs, starting with a clean slate on each task and only pulling in the information that it needs. We turned a human in the loop batch processing task into an AI creative workflow by splitting it into discrete steps and letting it kind of pick up where it left off. We gave the model a way to track progress. We built templates to make the results more consistent and reusable, and we recorded information about the database to help the AI model get smarter over time. I've had a ton of fun nailing down this workflow and testing it on all sorts of tasks. For my work, we're actually using it to do things similar to this where we're analyzing a huge amount of content and trying to figure out how to organize and categorize and and do all sorts of things, and it's it's doing really well. The ability to take notes, record what it's done, and just kind of start over and create its own context has been incredibly helpful with turning a workflow that works really well for one or two articles into something that works well across thousands. So thank you for your time. I really look forward to seeing what everyone does with Directus and Directus MCP going forward, especially as Directus MCP evolves, AI models get more capable, and just the overall core gets stronger. The future is gonna be fun.","61b97ecd-ad99-4e38-9799-fd3fbfe95e44",[],[],{"id":134,"number":135,"show":122,"year":136,"episodes":257},[138,139,140,141,142,143,144,145,146],{"id":145,"slug":259,"vimeo_id":260,"description":261,"tile":262,"length":200,"resources":8,"people":8,"episode_number":263,"published":154,"title":264,"video_transcript_html":265,"video_transcript_text":266,"content":8,"seo":267,"status":131,"episode_people":268,"recommendations":269,"season":270},"mcp-build-content-handling","1138950576","Wayne Eldridge talks through using the MCP for content handling.","fdc35657-28da-49fd-b023-ec3de32261c3",8,"Community MCP Build: Content Handling","\u003Cp>Speaker 0: My name is Wayne Eldridge. I run a company called Anamica in sunny Cape Town, South Africa. We build everything from mobile apps to custom business internal facing apps, primarily indirectus, and that's how we discovered directus at the back end. When the directors MCP server got announced, we were very excited. We've done a whole internal shift, from programming line by line to, obviously, tab completions to agentic, coding and agentic software development, with the human in the loop.\u003C/p>\u003Cp>So everything's still verified by a human, but we have a whole system of, feedback loops loops, error logging, everything like that. So when the the director's MCP server, arrived, we were in the process of porting our website from Webflow, to Next. Js. Webflow did have its own built in CMS. However, we needed the customization of Directus, and we were very familiar with the product.\u003C/p>\u003Cp>So when the Directus MCP server was announced, we saw a great opportunity to port all of our blogs and content, from the, web flow CMS, into directors, specifically using, autonomous agents. Typically, what we would do is self host direct us. However, you can use Directus Cloud, to initiate this as a general blog post engine itself. We used Next. Js in the front end and, obviously, Directus in the back end.\u003C/p>\u003Cp>We would use Cursor and Sonnet four, whatever model that you wish to select in Cursor to take the content from, Webflow, via the Webflow MCP server, use cursor and sonnet to connect to the director's MCP server, choose an updated, updated director's blog post, and associate that content to that blog post category. We then discovered that whilst we ported several of our blogs, we still had a whole bunch of links of interesting, tech articles that we stored in Notion, and we thought ourselves, you know, we could share these links and specific content extremely quickly with the director's MCP server. We dropped in one or two links into Cursor, asked the, agent in Cursor to write a blog post about it, and send it to the MCP server associating the blog post category to, these content of the article. This happened very quickly and very seamlessly between the directors, MCP server, our agent, and the links sitting locally on our machine. We then decided to port all our links of interesting tech that we store internally at Dynamic and, put them on the websites under our resource hub.\u003C/p>\u003Cp>We can also receive articles written by our copywriters, put them within our code base, and ask the director's MCP server to, post the content onto our website. And along with that, update the website, which is written in Next. Js with any SEO, requirements that has been picked up by the content of the article itself, or without leaving cursor, linked to a director's MCP, and directors either sitting on a self hosted instance or, if you wish, in the directors cloud.\u003C/p>","My name is Wayne Eldridge. I run a company called Anamica in sunny Cape Town, South Africa. We build everything from mobile apps to custom business internal facing apps, primarily indirectus, and that's how we discovered directus at the back end. When the directors MCP server got announced, we were very excited. We've done a whole internal shift, from programming line by line to, obviously, tab completions to agentic, coding and agentic software development, with the human in the loop. So everything's still verified by a human, but we have a whole system of, feedback loops loops, error logging, everything like that. So when the the director's MCP server, arrived, we were in the process of porting our website from Webflow, to Next. Js. Webflow did have its own built in CMS. However, we needed the customization of Directus, and we were very familiar with the product. So when the Directus MCP server was announced, we saw a great opportunity to port all of our blogs and content, from the, web flow CMS, into directors, specifically using, autonomous agents. Typically, what we would do is self host direct us. However, you can use Directus Cloud, to initiate this as a general blog post engine itself. We used Next. Js in the front end and, obviously, Directus in the back end. We would use Cursor and Sonnet four, whatever model that you wish to select in Cursor to take the content from, Webflow, via the Webflow MCP server, use cursor and sonnet to connect to the director's MCP server, choose an updated, updated director's blog post, and associate that content to that blog post category. We then discovered that whilst we ported several of our blogs, we still had a whole bunch of links of interesting, tech articles that we stored in Notion, and we thought ourselves, you know, we could share these links and specific content extremely quickly with the director's MCP server. We dropped in one or two links into Cursor, asked the, agent in Cursor to write a blog post about it, and send it to the MCP server associating the blog post category to, these content of the article. This happened very quickly and very seamlessly between the directors, MCP server, our agent, and the links sitting locally on our machine. We then decided to port all our links of interesting tech that we store internally at Dynamic and, put them on the websites under our resource hub. We can also receive articles written by our copywriters, put them within our code base, and ask the director's MCP server to, post the content onto our website. And along with that, update the website, which is written in Next. Js with any SEO, requirements that has been picked up by the content of the article itself, or without leaving cursor, linked to a director's MCP, and directors either sitting on a self hosted instance or, if you wish, in the directors cloud.","ee3f18af-3840-4cf7-bb81-794869789eb5",[],[],{"id":134,"number":135,"show":122,"year":136,"episodes":271},[138,139,140,141,142,143,144,145,146],{"id":146,"slug":273,"vimeo_id":274,"description":275,"tile":276,"length":170,"resources":8,"people":8,"episode_number":248,"published":154,"title":277,"video_transcript_html":278,"video_transcript_text":279,"content":8,"seo":280,"status":131,"episode_people":281,"recommendations":283,"season":284},"mcp-hackathon-challenges","1138984257","Bryant talks through the MCP Hackathon, giving details on the challenges, prizes and entry information.","e4f205b0-a2f5-41f3-9847-202fc5de6289","MCP Hackathon: Challenges, Prizes, and How to Enter","\u003Cp>Speaker 0: A few examples of what you can do with the Directus MCP. Now is your chance to actually build something with it, and we're gonna make it worth your while. So I'm here with the official announcement. We are running a Directus MCP hackathon for the next few weeks, and man, it's gonna be loads of fun. There are two challenges, and yes, you can enter both.\u003C/p>\u003Cp>Challenge number one is to build a cool Directus project using our native MCP. It's completely open ended. No rules, no limits. You can build a tool that solves a real problem or build something freaking wild. To enter, just provide a short overview and a quick five minute demo video showing us and the community how awesome it is.\u003C/p>\u003Cp>Challenge number two is to create a community template. You'll build a ready to go direct to starter that other developers can fire up right away. Think an AI powered blog or maybe a smart product catalog. Something that is helpful for the community and showcases what's possible when you build with the Directus MCP. All you need is a public GitHub repo with the template included.\u003C/p>\u003Cp>And, look, don't worry. We've got some tooling and a sample repo to make this process easy peasy for you. Winners in each challenge will get a cool thousand bucks USD, and even cooler than that is that every team member gets an eight bit dough retro mechanical keyboard. Yes, those are the ones that look like the old Nintendo, and yes, I will be jealous of you. Winners also get some sweet Directus swag.\u003C/p>\u003Cp>So get yourself registered at directus.i0/hackathons. Registering also gets you into our mentor hours sessions where you can ask questions and get help from other members of the core team and myself. Again, check out the official rules and submission details at directus.i0/hackathons. If you've been looking for an excuse to build something with AI and Directus, now's the time. I'm super excited to see what you build.\u003C/p>","A few examples of what you can do with the Directus MCP. Now is your chance to actually build something with it, and we're gonna make it worth your while. So I'm here with the official announcement. We are running a Directus MCP hackathon for the next few weeks, and man, it's gonna be loads of fun. There are two challenges, and yes, you can enter both. Challenge number one is to build a cool Directus project using our native MCP. It's completely open ended. No rules, no limits. You can build a tool that solves a real problem or build something freaking wild. To enter, just provide a short overview and a quick five minute demo video showing us and the community how awesome it is. Challenge number two is to create a community template. You'll build a ready to go direct to starter that other developers can fire up right away. Think an AI powered blog or maybe a smart product catalog. Something that is helpful for the community and showcases what's possible when you build with the Directus MCP. All you need is a public GitHub repo with the template included. And, look, don't worry. We've got some tooling and a sample repo to make this process easy peasy for you. Winners in each challenge will get a cool thousand bucks USD, and even cooler than that is that every team member gets an eight bit dough retro mechanical keyboard. Yes, those are the ones that look like the old Nintendo, and yes, I will be jealous of you. Winners also get some sweet Directus swag. So get yourself registered at directus.i0/hackathons. Registering also gets you into our mentor hours sessions where you can ask questions and get help from other members of the core team and myself. Again, check out the official rules and submission details at directus.i0/hackathons. If you've been looking for an excuse to build something with AI and Directus, now's the time. I'm super excited to see what you build.","4152664e-dc82-49fb-a1a2-54ca097c5c39",[282],"ee3c6dd7-e0a8-453e-99ae-ebc7fc869805",[],{"id":134,"number":135,"show":122,"year":136,"episodes":285},[138,139,140,141,142,143,144,145,146],{"reps":287},[288,344],{"name":289,"sdr":8,"link":290,"countries":291,"states":293},"John Daniels","https://meet.directus.io/meetings/john2144/john-contact-form-meeting",[292],"United States",[294,295,296,297,298,299,300,301,302,303,304,305,306,307,308,309,310,311,312,313,314,315,316,317,318,319,320,321,322,323,324,325,326,327,328,329,330,331,332,333,334,335,336,337,338,339,340,341,342,343],"Michigan","Indiana","Ohio","West Virginia","Kentucky","Virginia","Tennessee","North Carolina","South Carolina","Georgia","Florida","Alabama","Mississippi","New York","MI","IN","OH","WV","KY","VA","TN","NC","SC","GA","FL","AL","MS","NY","Connecticut","CT","Delaware","DE","Maine","ME","Maryland","MD","Massachusetts","MA","New Hampshire","NH","New Jersey","NJ","Pennsylvania","PA","Rhode Island","RI","Vermont","VT","Washington DC","DC",{"name":345,"link":346,"countries":347},"Michelle Riber","https://meetings.hubspot.com/mriber",[348,349,350,351,352,353,354,355,356,357,358,359,360,361,362,363,364,365,366,367,368,369,370,371,372,373,374,375,376,377,378,379,380,381,382,383,384,385,386,387,388,389,390,391,392,393,394,395,396,397,398,399,400,401,402,403,404,405,406,407,408,409,410,411,412,413,414,415,416,417,418,419,420,421,422,423,424,425,426,427,428,429,430,431,432,433,434,435,436,437,438,439,440,441,442,443,444,445,446,447,448,449,450,451,452,453,454,455,456,457,458,459,460,461,462,463,464,465,466,467,468,469,470,471,472,473,474,475,476,477,478,479,480,481,482,483,484,485,486,487,488,489,490,491,492,493,494,495,496,497,498,499,500,501,502,503,504,505,506,507,508,509,510,511,512,513,514,515,516,517,518,519,520,521,522,523,524,525,526,527,528,529,530,531,532,533,534,535,325,536,537],"Albania","ALB","Algeria","DZA","Andorra","AND","Angola","AGO","Austria","AUT","Belgium","BEL","Benin","BEN","Bosnia and Herzegovina","BIH","Botswana","BWA","Bulgaria","BGR","Burkina Faso","BFA","Burundi","BDI","Cameroon","CMR","Cape Verde","CPV","Central African Republic","CAF","Chad","TCD","Comoros","COM","Côte d'Ivoire","CIV","Croatia","HRV","Czech Republic","CZE","Democratic Republic of Congo","COD","Denmark","DNK","Djibouti","DJI","Egypt","EGY","Equatorial Guinea","GNQ","Eritrea","ERI","Estonia","EST","Eswatini","SWZ","Ethiopia","ETH","Finland","FIN","France","FRA","Gabon","GAB","Gambia","GMB","Ghana","GHA","Greece","GRC","Guinea","GIN","Guinea-Bissau","GNB","Hungary","HUN","Iceland","ISL","Ireland","IRL","Italy","ITA","Kenya","KEN","Latvia","LVA","Lesotho","LSO","Liberia","LBR","Libya","LBY","Liechtenstein","LIE","Lithuania","LTU","Luxembourg","LUX","Madagascar","MDG","Malawi","MWI","Mali","MLI","Malta","MLT","Mauritania","MRT","Mauritius","MUS","Moldova","MDA","Monaco","MCO","Montenegro","MNE","Morocco","MAR","Mozambique","MOZ","Namibia","NAM","Niger","NER","Nigeria","NGA","North Macedonia","MKD","Norway","NOR","Poland","POL","Portugal","PRT","Republic of Congo","COG","Romania","ROU","Rwanda","RWA","San Marino","SMR","São Tomé and Príncipe","STP","Senegal","SEN","Serbia","SRB","Seychelles","SYC","Sierra Leone","SLE","Slovakia","SVK","Slovenia","SVN","Somalia","SOM","South Africa","ZAF","South Sudan","SSD","Spain","ESP","Sudan","SDN","Sweden","SWE","Tanzania","TZA","Togo","TGO","Tunisia","TUN","Uganda","UGA","United Kingdom","GBR","Vatican City","VAT","Zambia","ZMB","Zimbabwe","ZWE","UK","Germany","Netherlands","Switzerland","CH","NL",1773850414674]