{
  "server": "elizaOS Development",
  "title": "elizaOS Development Discord - 2025-01-14",
  "date": 1736812800,
  "stats": {
    "totalMessages": 221,
    "totalUsers": 52
  },
  "categories": [
    {
      "channelId": "1320246527268098048",
      "channelName": "💬｜general",
      "summary": "# Analysis of Discord Chat in 💬｜general\n\n## 1. Summary\nThe chat primarily focused on technical issues and implementations related to the Eliza agent platform. Key discussions included:\n\n- Twitter client integration issues, particularly with formatting problems in tweets where numbered lists were appearing incorrectly. A fix was proposed by adding regex replacements to remove numbered list markers.\n- Custom plugin development for wallet tracking functionality, with KeyesCode.eth sharing their implementation of an agent that follows ETH and BASE wallets.\n- Questions about implementing automated actions through the Twitter client, specifically how to make an agent post content from custom actions at regular intervals.\n- Knowledge base implementation challenges, with users discussing how to make agents consistently respond with specific information like links.\n- Database integration issues, particularly with Supabase adapter changes and knowledge table writing problems.\n- Node version compatibility concerns, with warnings about Node 23.5.0 vs the expected 23.3.0.\n- Telegram client connection problems on Linux environments.\n\n## 2. FAQ\nQ: How do I make my agent automatically post content from custom actions at regular intervals? (asked by KeyesCode.eth) A: You need to create a provider and pass that provider into the Twitter post template, then amend the template to explain to the agent that you want the provider data in the tweet every time. (answered by R0am)\nQ: How can I make my agent answer with a specific link when asked? (asked by bendermind) A: Add the links to the 'system' prompt in char.json. (answered by bendermind)\nQ: Is it possible for the agent to join an existing space rather than starting its own? (asked by pluto) A: Yes, agent-twitter-client can do that, but client-twitter doesn't have the needed code to take advantage of the join spaces functionality yet. (answered by sayonara)\nQ: Is there a way to give the agent knowledge of things like docs on a URL? (asked by pluto) A: Not currently. The ragKnowledge is an initial implementation. For advanced RAG solutions, a custom plugin would be needed. If you get the docs in markdown or txt format, you can use those. (answered by ꧁Ninja_Dev꧂)\nQ: Why is my agent not responding to 'gm' when tagged? (asked by bendermind) A: Unanswered\nQ: Can I increase the threshold in the knowledge folder system to keep my agent from hallucinating? (asked by Mad) A: Unanswered\nQ: Is it possible to integrate an Eliza agent into a browser extension? (asked by eeee) A: Unanswered\nQ: Where can we find v2 core code? (asked by Mike D.) A: I don't believe it's out yet. (answered by ꧁Ninja_Dev꧂)\nQ: What's the latest stable version of the Eliza repo? (asked by ICJR) A: v0.1.8+build.1 should be more stable. It's basically v0.1.8.1 but semver doesn't allow that. (answered by Odilitime)\n\n## 3. Help Interactions\nHelper: sayonara | Helpee: 4n7m4n | Context: Issues with tweet formatting after 1.8+ update causing numbered lines and quotes | Resolution: Identified PR that might have caused the issue, 4n7m4n created a fix with regex replacements and submitted PR #2299\nHelper: R0am | Helpee: KeyesCode.eth | Context: How to make Twitter client post content from custom actions automatically | Resolution: Suggested creating a provider and passing it to the Twitter post template\nHelper: sayonara | Helpee: pluto | Context: Understanding how to make an agent join existing Twitter spaces | Resolution: Provided links to relevant code and explained current limitations\nHelper: ꧁Ninja_Dev꧂ | Helpee: pluto | Context: Questions about giving agent knowledge from URLs | Resolution: Explained current limitations of ragKnowledge and suggested using markdown/txt formats\nHelper: Mike D. | Helpee: Community | Context: Getting Eliza running on arm64 architecture | Resolution: Successfully implemented and shared the solution on Twitter\n\n## 4. Action Items\nType: Technical | Description: Fix tweet formatting issues by adding regex to remove numbered list markers | Mentioned By: 4n7m4n\nType: Technical | Description: Implement functionality for agents to join existing Twitter spaces | Mentioned By: pluto\nType: Technical | Description: Investigate why knowledge is not being written to the Supabase DB knowledge table | Mentioned By: 4n7m4n\nType: Technical | Description: Fix Telegram client connection issues on Linux environments | Mentioned By: Founder\nType: Feature | Description: Add capability for more advanced RAG knowledge implementation, including URL parsing | Mentioned By: pluto\nType: Feature | Description: Implement action processing for Twitter likes | Mentioned By: raggamuffn\nType: Documentation | Description: Improve documentation on setting up target users for Twitter | Mentioned By: flow_moca\nType: Documentation | Description: Add more documentation on GOAT SDK | Mentioned By: validsyntax\nType: Documentation | Description: Update node version requirements in documentation | Mentioned By: 4n7m4n",
      "messageCount": 152,
      "userCount": 35
    },
    {
      "channelId": "1324098367416172665",
      "channelName": "📮｜feedback",
      "summary": "# Analysis of Discord Chat in 📮｜feedback Channel\n\n## 1. Summary\nThe discussion centers on package management concerns in the project repository. Users are experiencing frustration with the current monolithic structure where all adapters, clients, and packages are bundled together. This causes several issues:\n- Unnecessary imports remain when updating to new releases\n- High memory consumption (6.7GB RAM for running a single agent)\n- Large node_modules size (3GB), primarily due to llama-cpp and duplicate onnxruntime versions from the transformers package\n- Concerns about cost-efficiency when deploying to cloud environments (AWS/GCP)\n\nThe main proposal is to segregate packages/adapters/clients from the main repository and move to a manual package installation approach, allowing users to include only what they need.\n\n## 2. FAQ\nQ: Why is the node_modules size so large? (implied by mangu) A: It's 3GB due to llama-cpp and two versions of onnxruntime because of the transformers package version (answered by Cesar Rodriguez)\n\n## 3. Help Interactions\nHelper: Cesar Rodriguez | Helpee: mangu | Context: Explaining why node_modules is so large | Resolution: Identified specific packages causing bloat (llama-cpp and duplicate onnxruntime versions)\n\n## 4. Action Items\nType: Technical | Description: Segregate packages/adapters/clients from the main repository | Mentioned By: mangu\nType: Technical | Description: Implement manual package installation to reduce unnecessary dependencies | Mentioned By: mangu\nType: Technical | Description: Optimize memory usage to reduce the 6.7GB RAM requirement for running one agent | Mentioned By: mangu\nType: Technical | Description: Address duplicate onnxruntime versions from transformers package | Mentioned By: Cesar Rodriguez",
      "messageCount": 3,
      "userCount": 3
    },
    {
      "channelId": "1324089429727514674",
      "channelName": "🤖｜agent-dev-school",
      "summary": "# Analysis of \"🤖｜agent-dev-school\" Discord Channel\n\n## 1. Summary\nThe channel discussions primarily focus on ElizaOS development, with users seeking guidance on setup, implementation, and extending functionality. Several users are working on creating agentic applications, with specific interest in building personalized agents using vector databases and character files. There are questions about repository structure, API integration, and plugin development. Users share links to resources including GitHub repositories (elizaOS/characterfile, elizaOS/eliza) and announcements about ElizaOS v2. One user describes wanting to create an agent with more advanced capabilities beyond basic chatbot functionality, expressing frustration with the complexity of the codebase. Another user is working on creating an agentic version of themselves using their written articles and chat history. The channel also contains announcements about AI agent development opportunities and resources.\n\n## 2. FAQ\nQ: Are there any more advanced tutorials for eliza for instance about editing {{actions}} to make the bot execute an action? (asked by shmaven) A: Unanswered\nQ: Is it possible to run the agent from a different repository? (asked by chui) A: Unanswered\nQ: Where can we find the new v2 core code? (asked by Mike D.) A: Mike D. later shared links to a GitHub PR and Twitter post about ElizaOS v2\nQ: What's the most stable branch to build off of right now? (asked by saccharinedreams) A: Unanswered\nQ: Which repo should I open a pull request to for my plugin to be available in the Eliza OS? (asked by Samidha) A: Cesar Rodriguez shared a link to https://github.com/elizaOS/eliza\n\n## 3. Help Interactions\nHelper: CheddarQueso 🧀 | Helpee: sirprincess | Context: API key setup and initialization issues | Resolution: sirprincess figured out that the system makes calls to Anthropic once chatting starts, despite initializing LlamaService at startup\nHelper: epilogg | Helpee: (implied previous question about character files) | Context: Sharing resource for character file creation | Resolution: Shared link to tweets2character tool in GitHub repository\nHelper: Mike D. | Helpee: Channel members | Context: Finding ElizaOS v2 code | Resolution: Shared links to GitHub PR and Twitter post about ElizaOS v2\nHelper: Cesar Rodriguez | Helpee: Samidha | Context: Where to submit plugin PR | Resolution: Shared link to the main ElizaOS repository\n\n## 4. Action Items\nTechnical: Create an agentic version of self using articles and chat history with vector DB | Description: Implement personal agent that can pass a Turing test with domain experts | Mentioned By: Rabbidfly\nTechnical: Implement advanced action capabilities for ElizaOS agents | Description: Enable agents to execute tasks while maintaining context of memory and goals | Mentioned By: shmaven\nDocumentation: Need advanced tutorials for editing {{actions}} in ElizaOS | Description: Documentation on how to implement agentic properties beyond basic chatbot functionality | Mentioned By: shmaven\nFeature: Improve agent's ability to sound more human-like | Description: Enhanced natural language generation capabilities | Mentioned By: shmaven",
      "messageCount": 23,
      "userCount": 13
    },
    {
      "channelId": "1323745969115893780",
      "channelName": "📥｜pull-requests",
      "summary": "# Analysis of \"📥｜pull-requests\" Discord Channel\n\n## 1. Summary\nThe channel primarily focuses on pull request discussions for the ElizaOS project. Key technical discussions centered around:\n\n1) Solana plugin improvements: A user (suns) enhanced the legacy Solana plugin to properly handle native SOL and SPL transfers, submitting PR #2340. There was confusion about two Solana plugins existing simultaneously (plugin-solana and plugin-solana-agentkit).\n\n2) ModelProvider duplication issue: Cesar Rodriguez identified a potential issue where modelProvider exists in both AgentRuntime and character objects, causing unexpected behavior when configurations conflict. He suggested removing the redundant modelProvider from runtime since it's mandatory in character.\n\n3) Supabase implementation: AIFlow.ML proposed reviving PR #1800 as an alternative to the current Supabase implementation which has received multiple fix attempts. Odilitime agreed to review this approach.\n\n4) Agent scaling: A brief discussion revealed Odilitime runs 350 agents on a single EPYC processor instance, contrasting with N3rdStorm's preference for single-agent deployments.\n\n5) Several PRs were mentioned for review including #2275, #49, #1800, #2335, #2338, and #2149, covering various improvements to ElizaOS.\n\n## 2. FAQ\nQ: Why does AgentRuntime have a modelProvider if character has a modelProvider? (asked by Cesar Rodriguez) A: Because agentruntime is basically the instance of a character, they're pretty much the same. It's an artifact of optimizing for a single character but allowing multiple agents. (answered by Odilitime)\nQ: How many agents can be run on one instance? (implied by N3rdStorm) A: I run 350 agents on one instance (answered by Odilitime)\nQ: Why are there two Solana plugins (plugin-solana and plugin-solana-agentkit)? (asked by suns) A: Agent-kit got its own plugin for now (answered by Odilitime)\n\n## 3. Help Interactions\nHelper: ꧁Ninja_Dev꧂ | Helpee: suns | Context: How to get Solana plugin improvements merged into main repo | Resolution: Advised to submit a PR against the develop branch for review and merging\nHelper: Odilitime | Helpee: Cesar Rodriguez | Context: Confusion about duplicate modelProvider in AgentRuntime and character | Resolution: Explained it's an artifact of optimization, noting .env is not recommended for multiple agents but useful for maintaining defaults\n\n## 4. Action Items\nTechnical: Review and potentially merge improved Solana plugin that handles native SOL and SPL transfers | Description: PR #2340 fixes functionality in legacy Solana plugin | Mentioned By: suns\nTechnical: Consider removing modelProvider from runtime since it's mandatory in character | Description: Refactor to prevent unexpected behavior from duplicate configuration | Mentioned By: Cesar Rodriguez\nTechnical: Review PR for autonomous agent integration in Discord and Telegram | Description: PRs #2335 & #2338 created for more ingrained autonomous agents | Mentioned By: ꧁Ninja_Dev꧂\nTechnical: Consider reviving PR #1800 as alternative to problematic Supabase implementation | Description: Previous implementation was working and tested | Mentioned By: AIFlow.ML @ ElizaOS\nTechnical: Review PR #2149 which appears ready for merging | Description: Mentioned as \"looks good to go\" | Mentioned By: Santi\nTechnical: Review PR #2275 for potential integration | Description: Requested for review | Mentioned By: AIFlow.ML @ ElizaOS\nTechnical: Test patch in PR #49 for agent-twitter-client | Description: Requested for testing | Mentioned By: sayonara",
      "messageCount": 43,
      "userCount": 10
    }
  ]
}