{
  "server": "Hyperfy",
  "title": "Hyperfy Discord - 2025-08-20",
  "date": 1755648000,
  "stats": {
    "totalMessages": 243,
    "totalUsers": 27
  },
  "categories": [
    {
      "channelId": "994775534733115412",
      "channelName": "💻│developers",
      "summary": "# Analysis of 💻│developers Discord Chat\n\n## 1. Summary\nThe discussion primarily focused on app architecture in Hyperfy, with extensive conversation about how to structure apps for game development. Lastraum asked about spawning GLB models in world space, leading to a detailed explanation from ash about Hyperfy's app-based architecture. Each app combines a GLB model, script, and metadata, with different approaches for handling multiple game objects. Ash explained that while everything must be an app, developers can structure their code in various ways - either as many small specialized apps or fewer apps with more complex functionality. The chat also covered AI-generated content in Hyperfy's \"vibe world\" where users can create and edit 3D objects through text commands, with ash noting memory leak fixes. A significant technical development was shared by mrtzhckr who demonstrated Gaussian Splat rendering in Hyperfy, with performance considerations for different platforms. Other topics included local development workflows, testing on Quest devices using port forwarding, and UI/UX considerations for avatar selection and pointer lock behavior.\n\n## 2. FAQ\nQ: How can an app spawn a GLB in world space? (asked by Lastraum - theforgecore.xyz) A: You can use world.add(mesh) which will remove it from its previous parent. (answered by ash)\nQ: If I have 50 items in a \"kitchen\", do they all need to be their own individual app? (asked by Lastraum - theforgecore.xyz) A: Everything is apps, but you can structure them in different ways - either as many specialized apps or fewer apps with more complex functionality. (answered by ash)\nQ: Does the Image Asset only work when live and not when running locally? (asked by TheMattEmpire) A: Local and live shouldn't be any different. (answered by ash)\nQ: How can I test networked stuff locally without SSL certs on Quest? (asked by mrtzhckr) A: Use port forwarding with Meta Quest Developer Hub: \"adb reverse tcp:3000 tcp:3000\" so when you visit localhost:3000 on Quest it's the same as on your computer. (answered by ash)\nQ: What model does the AI in vibe world use? (asked by vox) A: Sonnet 4. (answered by ash)\nQ: What's the performance like on Gaussian Splats? (asked by ash) A: Performance depends on file size - under 50MB works well on desktop, 5-8MB limit for Quest before performance drops. (answered by mrtzhckr)\nQ: Is there a way to force FPV or detect when in FPV/3PV? (asked by Lastraum - theforgecore.xyz) A: Neither yet. (answered by ash)\n\n## 3. Help Interactions\nHelper: ash | Helpee: Lastraum - theforgecore.xyz | Context: How to structure apps for a game with multiple objects like harvestable resources | Resolution: Explained different approaches - either one app per category with shared code or a game manager app that coordinates multiple specialized apps\nHelper: ash | Helpee: mrtzhckr | Context: Testing networked features on Quest without SSL | Resolution: Provided ADB command for port forwarding to avoid SSL requirements\nHelper: ash | Helpee: TheMattEmpire | Context: Image Asset not working in local development | Resolution: Suggested refreshing, which resolved the issue\nHelper: Shiffty | Helpee: Lastraum - theforgecore.xyz | Context: How to spawn a GLB in world space | Resolution: Explained that after cloning a node, you need to app.remove it and world.add it\nHelper: ash | Helpee: vox | Context: Fullscreen toggle issue when switching mouselock/free mouse | Resolution: Added support for blockquote key (`) to toggle pointer lock without exiting fullscreen\n\n## 4. Action Items\nTechnical: Fix memory leak in AI-generated content feature | Description: Material cache on primitives causing memory buildup | Mentioned By: ash\nTechnical: Implement Gaussian Splat rendering support | Description: Optimize for performance across desktop and mobile platforms | Mentioned By: mrtzhckr\nTechnical: Add alternative button for pointer lock toggle | Description: Use blockquote key (`) to toggle pointer lock without exiting fullscreen | Mentioned By: ash\nFeature: Add ability to force first-person view or detect view mode | Description: Allow developers to force or detect FPV/TPV camera modes | Mentioned By: Lastraum - theforgecore.xyz\nFeature: Add \"F\" key shortcut for fullscreen toggle | Description: Implement standard fullscreen shortcut for better user experience | Mentioned By: vox\nDocumentation: Create guide on app architecture patterns | Description: Document different approaches to structuring apps for game development | Mentioned By: ash",
      "messageCount": 164,
      "userCount": 16
    },
    {
      "channelId": "958209074045026327",
      "channelName": "⚡│general",
      "summary": "# Discord Chat Analysis for \"⚡│general\"\n\n## 1. Summary:\nThe chat primarily revolves around Hyperfy's new AI-powered world creation feature called \"Vibe.\" Users are sharing links to social media posts about the technology and expressing excitement. Ash provided basic instructions for using the AI commands in Vibe, specifically `/create a bookshelf` to generate objects and `/edit add some books` to modify existing objects. The feature is described as being in \"super alpha\" stage. There was a brief mention of an error with Anthropic's API being temporarily overloaded, causing the Vibe world AI to go down. One user suggested changing the generative model to Grok for potential viral growth. A user named SEBGA reported issues with the platform, including losing chat functionality and being unable to jump after refreshing.\n\n## 2. FAQ:\nQ: What skills should I have to try it, have you any tips to share? (asked by 🆂🅴🅱🅶🅰) A: Unanswered\nQ: What are the asset checked by the role bot? it say i don't have any but I have 3 lands on the wallet scanned. (asked by 🆂🅴🅱🅶🅰) A: Unanswered\nQ: How do I use Vibe? (implied by 🆂🅴🅱🅶🅰) A: Use `/create a bookshelf` to create something with AI where you are looking and `/edit add some books` to edit something you are pointing at (answered by ash)\n\n## 3. Help Interactions:\nHelper: ash | Helpee: 🆂🅴🅱🅶🅰 | Context: User didn't know how to use Vibe | Resolution: Provided link to vibe.hyperfy.xyz and basic commands for creating and editing objects with AI\nHelper: ash | Helpee: Channel users | Context: Vibe AI service outage | Resolution: Informed users that the Vibe world AI was down due to Anthropic's API being temporarily overloaded\n\n## 4. Action Items:\nFeature: Consider changing the generative model to Grok for potential viral growth | Mentioned By: MetaMike\nTechnical: Fix issue where users can't jump after refreshing the page | Mentioned By: 🆂🅴🅱🅶🅰\nTechnical: Address role bot not recognizing owned land assets | Mentioned By: 🆂🅴🅱🅶🅰",
      "messageCount": 26,
      "userCount": 11
    },
    {
      "channelId": "1326789867312775290",
      "channelName": "🪙│hyper",
      "summary": "# Analysis of Discord Chat in \"🪙│hyper\" Channel\n\n## 1. Summary\nThe chat segment primarily revolves around interpersonal conflict rather than technical discussions. There appears to be tension regarding a former contributor named \"saori\" who has created a new account to post in the channel. Community members express frustration about this individual's behavior, with ash mentioning that they improved the \"primitives\" component to be \"100x more performant.\" The conversation suggests there was a history of collaboration issues where work would be done jointly but credit would later be claimed by saori. PabloFM | Numinia posts a supportive message for ash, referencing the GPL license and encouraging focus on the community's work rather than the conflict. No specific technical implementations, decisions, or problem-solving discussions are present in this chat segment.\n\n## 2. FAQ\nQ: None present in this chat segment.\n\n## 3. Help Interactions\nNone present in this chat segment.\n\n## 4. Action Items\nNone explicitly mentioned in this chat segment.",
      "messageCount": 23,
      "userCount": 7
    },
    {
      "channelId": "1031058655581323324",
      "channelName": "🧊│3d-design",
      "summary": "# Discord Chat Analysis for 🧊│3d-design\n\n## 1. Summary\nThe chat segment focuses on 3D design performance testing in Hyperfy. TheMattEmpire is experimenting with instancing to test performance limits, noting that it works well in localized areas despite being time-consuming without multiselect functionality. They're using emission to simulate lighting while maintaining shadows for objects inside buildings by disabling shadows on the roof. Their goal is to incrementally add content to identify performance thresholds. The conversation briefly shifts to streaming capabilities in Hyperfy, with ash suggesting that .m3u8 streams might work (as they did in v1), and .hyp shaman confirming they use bloomers.tv to generate m3u8 links that work well in the platform.\n\n## 2. FAQ\nQ: Does Hyperfy have streaming into spaces yet, or just MP4 support? (asked by TheMattEmpire) A: Live streaming should work if you use .m3u8 streams from OBS or similar platforms (answered by ash)\nQ: How can I get an m3u8 link for streaming? (implied by context) A: bloomers.tv provides m3u8 links easily and works well (answered by .hyp shaman)\n\n## 3. Help Interactions\nHelper: ash | Helpee: TheMattEmpire | Context: Inquiring about streaming capabilities in Hyperfy | Resolution: Suggested that .m3u8 streams should work as they did in v1\nHelper: .hyp shaman | Helpee: TheMattEmpire | Context: Needing a way to get m3u8 links for streaming | Resolution: Recommended bloomers.tv as an easy solution for generating m3u8 links\n\n## 4. Action Items\nTechnical: Implement multiselect functionality for more efficient instancing | Description: Current instancing process is time-consuming without multiselect | Mentioned By: TheMattEmpire\nTechnical: Continue testing performance limits by incrementally adding content | Description: Determine at what point frame rate begins to drop | Mentioned By: TheMattEmpire\nFeature: Support for multiselect in the editor | Description: Would make instancing workflow much faster | Mentioned By: TheMattEmpire",
      "messageCount": 12,
      "userCount": 4
    },
    {
      "channelId": "999870741069692958",
      "channelName": "🌆│share",
      "summary": "The chat segment is extremely brief, containing only a single message from a user named Rick. The message indicates that Rick shared a tweet from the Twitter user \"em0tionull\" that was originally posted by Discord user with ID 189071158669213696. The tweet was shared approximately 1 minute before the chat log was captured. There is no technical discussion, problem-solving, or implementation details in this brief exchange.",
      "messageCount": 1,
      "userCount": 1
    },
    {
      "channelId": "1022421054582423562",
      "channelName": "🐞│issues",
      "summary": "# Discord Chat Analysis\n\n## 1. Summary\nThe discussion in the \"🐞│issues\" channel focused on improving the precision of object positioning in worlds. User \"vox\" highlighted that the current input fields only allow positioning down to 0.1 meters, which lacks the necessary precision for detailed work. While the gizmo tool allows for finer adjustments, the numerical input fields needed improvement. Developer \"ash\" responded quickly, implementing a solution in the latest dev version that allows for different step sizes when using keyboard shortcuts: Alt+arrow keys for 0.01 increments (fine), regular arrow keys for 0.1 increments (standard), and Shift+arrow keys for 1.0 increments (large). Similar functionality was added for rotation with appropriate degree values. User \"𝕽𝖔𝖚𝖘𝖙𝖆𝖓\" reported difficulty using these new controls on Mac, potentially due to the Alt key mapping to Option, and requested clarification.\n\n## 2. FAQ\nQ: Can we have finer steps (0.01) when arranging objects in worlds? (asked by vox) A: Yes, will add it to the input fields (answered by ash)\nQ: How do the new positioning controls work? (asked by 𝕽𝖔𝖚𝖘𝖙𝖆𝖓) A: Unanswered\n\n## 3. Help Interactions\nHelper: ash | Helpee: vox | Context: Need for finer positioning controls in input fields | Resolution: Implemented keyboard shortcuts for different increment sizes (0.01, 0.1, 1.0) in latest dev version\n\n## 4. Action Items\nType: Technical | Description: Investigate Mac compatibility issue with Alt key shortcuts for fine positioning | Mentioned By: 𝕽𝖔𝖚𝖘𝖙𝖆𝖓\nType: Documentation | Description: Create documentation for the new keyboard shortcuts for object positioning | Mentioned By: ash\nType: Technical | Description: Fix input field precision to allow manual entry of values with 0.01 precision | Mentioned By: vox",
      "messageCount": 7,
      "userCount": 3
    },
    {
      "channelId": "1332108186676891649",
      "channelName": "🏗│infra",
      "summary": "# Discord Chat Analysis for 🏗│infra\n\n## 1. Summary\nThe chat segment is brief and focuses on users hitting their monthly API usage limits. User \".hyp shaman\" mentions reaching their monthly limit, which appears to be 1GB on the free tier. They note they've been working on \"elizza\" locally and had been sharing their API key across multiple worlds (3 worlds plus 1 local). User \"ash\" reports a similar experience of hitting limits on the same day, suggesting the service provider may have fixed their usage calculation. The conversation indicates that various services are reducing their usage limits.\n\n## 2. FAQ\nQ: How much is the current plan? (asked by vox) A: 1GB a month for free (answered by .hyp shaman)\n\n## 3. Help Interactions\nNo significant help interactions were present in this chat segment.\n\n## 4. Action Items\nTechnical: Consider implementing better API key management to avoid sharing across multiple worlds | Mentioned By: .hyp shaman\nTechnical: Monitor API usage more closely as limits appear to be enforced more strictly now | Mentioned By: ash",
      "messageCount": 10,
      "userCount": 3
    }
  ]
}