# elizaOS Discord - 2025-02-12 **Date: February 12, 2025** ## Overall Discussion Highlights ### ElizaOS Development Updates - **ElizaOS V2 Development**: Plugins have been moved to separate repositories to enable more permissionless and scalable plugin development. Migration effort from V1 to V2 is estimated at 1-5 on a scale of 10. - **Launchpad/Marketplace**: Technical infrastructure is ready but undergoing final audits before launch. - **DegenAI Trading V2**: Now live with a new sentiment/data layer that integrates social signals from Twitter and Telegram to refine trading decisions. The system made its first purchase of $POPCAT. - **Autonomous Investor**: Trust marketplace is live and generating signals from Telegram, with Discord integration expected within a week. ### Technical Discussions - **Node.js Compatibility**: Version 23+ is recommended for ElizaOS deployments to avoid dependency issues. - **API Access**: Several users discovered that API endpoints should be accessed via port 3000 instead of 5173, with endpoints like localhost:3000/agents. - **Database Storage**: Confirmation that the SQL database stores agent tweets for memory persistence, accessible in agent/data/db.sqlite in the memories table. - **Model Configuration**: Users discussed how to configure different model providers (OpenAI, TogetherAI) and specify which models to use. - **Hosting Requirements**: 4-8GB RAM recommended for hosting ElizaOS agents, with various VPS providers suggested. ### AI-TV Show Development - **News Aggregator Issues**: The team is facing challenges with poor-quality or mismatched images for news segments. - **Automation vs. Curation**: Debate between maintaining full automation versus introducing manual curation steps to improve quality. - **Technical Solutions**: Suggestions included creating generic "stinger" images for each news category and adjusting the episode generator to use these instead of aggregator image URLs. ### Branding and Tokenomics - **Brand Consolidation**: Discussion about whether to consolidate the ElizaOS (blue theme, professional) and ai16zdao (orange theme, crypto culture) accounts into a single brand identity, with most partners favoring consolidation. - **Tokenomics Model**: Current plan involves a dual pool system: SOL:AT (agent token) as primary and AI16Z:AT as a secondary pool that captures ecosystem fees for AI16Z buybacks. Some community members advocated for directly pairing agent tokens with AI16Z (similar to Virtuals and Arc models). ## Key Questions & Answers ### Development & Technical - **Q**: Does the SQL database store tweets the agent makes for memory? **A**: Yes, you can check them locally inside agent/data/db.sqlite in the memories table. - **Q**: How do I set up my agent to use OpenAI? **A**: Add your API key to .env and set "modelProvider": "openai" in your character.json file. - **Q**: Is it possible to create threads in X/Twitter using Eliza? **A**: Yes, check the utils.ts inside client-twitter and look for buildConversationThread method. - **Q**: How do I make my agent post to Twitter at regular intervals? **A**: Set Twitter credentials in .env and add "twitter" to the clients in your character.json. - **Q**: What are the recommended specs for hosting an Eliza agent? **A**: 4-8GB RAM. - **Q**: How do I call the Eliza Agent remotely? **A**: Use port 3000 instead of 5173, with endpoints like localhost:3000/agents. - **Q**: If I want to start building my own AI agent leveraging ElizaOS, is it better to wait for v2 or start now? **A**: Start now and migrate later, migration effort would be 1-5 on a scale of 10. ### DegenAI & Trading - **Q**: How is DegenAI different from quantitative trading bots and script-based bots? **A**: Quant-based bots use predefined mathematical algorithms, while script-based bots follow simple rules. DegenAI's new trading backend can implement any strategy while also incorporating sentiment analysis from social media to inform trading decisions. - **Q**: What are your thoughts on Arc launching a new AI agent that seems to be an enhanced version of DegenAI? **A**: Their tech is rust-based targeting specific users. There's overlap but it's not exactly the same - it's a platform, not an autonomously posting/trading agent. We can learn from their successes and failures. ### Branding & Tokenomics - **Q**: What's the reasoning for having separated brands? **A**: ElizaOS = blue theme, github org, professional front for partnerships; ai16zdao = investment dao, AI memecoin traders/VC, crypto culture. - **Q**: What is the current plan for token pairing? **A**: Dual pool system with SOL:AT as primary and AI16Z:AT as secondary pool. - **Q**: Are we following the same model as Virtuals and Arc? **A**: No, we're using a dual pool model that captures ecosystem fees to drive AI16Z buybacks. ## Community Help & Collaboration ### Technical Support - **Jox** helped **BowTiedBlueFin** resolve API call issues by suggesting using port 3000 instead of 5173 for API calls. - **DEVDARK** assisted **A ïert ꧂| ViP** with forwarding tweets to Discord/Telegram channels by providing code examples for accessing Discord client and sending messages to channels. - **elizaos-bridge-odi** helped **[elizaos] ** set up Twitter integration with Eliza by explaining how to configure Twitter credentials and add Twitter to clients in character.json. - **Odilitime** guided **Slise** on specifying which model to use with TogetherAI by directing them to edit the models.ts file. - **Derby** helped **minco** resolve a Docker build error with tokenizers module by suggesting installing wget and using node:23-slim in Dockerfile. ### Project Collaboration - **boom** created a News Curator tool (github.com/gm3/news-browser) to visualize and potentially curate the JSON data from the news aggregator. - **SM Sith Lord** provided a revised show-config.json with updated prompts to improve generation quality for the AI-TV show. - **Odilitime** shared a detailed analysis of $POPCAT from Agent Scarlett showing metrics, volume trends, and social sentiment as an example of AI analysis capabilities. ### Community Guidance - **jin** clarified the distinction between ElizaOS and ai16zdao brands for **flockaflame** and initiated a discussion about potential consolidation. - **rhota** provided a detailed explanation to **Red - X-Ware.v0** about how DegenAI differs from traditional trading bots and its sentiment analysis capabilities. - **witch** advised **Whimsical** to start building on ElizaOS now rather than waiting for V2, explaining that migration would be relatively easy. ## Action Items ### Technical - Implement dual pool system with SOL:AT primary and AI16Z:AT secondary (mentioned by jin) - Complete ElizaOS V2 development with plugins moved to separate repositories (mentioned by accelxr) - Finish Discord integration for Autonomous Investor trust marketplace (mentioned by accelxr) - Complete final audits for launchpad/marketplace (mentioned by accelxr) - Update Gaianet plugin to support API key (mentioned by Slise) - Fix Discord.js opus module installation issues (mentioned by [elizaos] ) - Implement user identification in client (mentioned by boolkeys) - Optimize token usage for long-running Twitter bots (mentioned by passer) - Add support for WebSocket connections (mentioned by ualp.) - Fix Docker environment variable loading (mentioned by BowTiedBlueFin) - Create generic "stinger" images for each news category (mentioned by SM Sith Lord) - Adjust episode generator to use category stingers instead of aggregator image URLs (mentioned by SM Sith Lord) - Add a fallback generic image for unexpected news categories (mentioned by SM Sith Lord) - Fix market data accuracy in the aggregator (mentioned by boom) ### Documentation - Create guide for remote API interaction (mentioned by BowTiedBlueFin) - Update Twitter client documentation (mentioned by amit) - Deliver product roadmap by Friday (mentioned by accelxr) - Create documentation for accelerator program ahead of ETHDenver (mentioned by accelxr) - Define a stable JSON specification for the news aggregator (mentioned by boom) - Create flowchart/infographic explaining DegenAI's architecture and functionality (mentioned by rhota) - Update tokenomics documentation with final model (mentioned by 阿拉斯加的狗🔯) ### Features - Add support for Facebook and Instagram posting (mentioned by POPPP) - Add user-specific memory to client (mentioned by cryptoAYA) - Consolidate social media accounts based on partner poll results (mentioned by jin and accelxr) - Launch Clank Tank show (mentioned by jin) - Implement tokenomics for the marketplace (mentioned by accelxr) - Develop accelerator program for agent projects (mentioned by accelxr) - Consider implementing a Twitter-focused show with embedded web browser for tweets (mentioned by SM Sith Lord) - Focus on market sentiment rather than specific numbers for AI-TV (mentioned by boom) - Purchase stinger images/videos for news categories (mentioned by jin) - Enhance sentiment analysis layer to process data from Twitter, Telegram, and other sources (mentioned by rhota) - Develop trading strategies that can adapt to market conditions (mentioned by Curtisdonthurtus)