# elizaOS Discord - 2025-03-19

## Overall Discussion Highlights

### ElizaOS v2 Development Status
- The v2 beta has recently gone live but remains somewhat unstable, with developers actively fixing bugs
- Beta phase is expected to last approximately two weeks according to rhota
- Complete refactoring has significantly improved developer experience with new features including:
  - GUI and Tauri app
  - CLI improvements
  - In-browser editor
  - Simplified API that eliminates character files in favor of runtime configurations
- The team is about to publish the next beta version and has fixed most startup issues
- Multiple repositories have been merged into a single ElizaOS core repo

### Technical Challenges & Solutions
- Users are experiencing various issues in the transition from v0.25.x to v1.0.0-beta/alpha versions
- RAG knowledge implementation is a common pain point, with consensus that placing files in `characters/knowledge/agent_knowledge/` works best
- Plugin compatibility issues are prevalent, especially with plugin-sql
- Several users report issues with CLI installation and startup process for newer versions
- Some users are working on adapters for databases like Supabase and MongoDB

### Documentation & Visual Assets
- Jin is leading efforts to create visual assets for ElizaOS documentation using AI image generation tools
- Documentation structure includes entities, components, actions, evaluators, providers, and databases
- Jin is working on improving navigation with mermaid flow charts
- Chinese community members requested clearer onboarding guides for developers

### Tokenomics & DAO Organization
- Discussions about potential token utility models for ai16z, including:
  - Using tokens as a commodity for paying for compute
  - Stake-weighted voting system for securing new Elizas/plugins
  - Agent registration system using public keys
- Plans for multi-chain support for the ElizaOS marketplace, with interest in BNB Chain integration
- Implementation of a structured social media amplification process
- Development of AI agents for governance with connections to Hats Protocol
- Collaboration with Agora to explore agent-based governance

### Spartan/DegenAI Integration
- The team is working on making Spartan functional within the new v2 architecture
- DegenAI's technology had to be moved entirely to the v2 stack
- Current priority is getting open-source functionality working in v2 and deploying Spartan to communities
- The spartan_holders channel was made private to give token holders exclusive access to the full Spartan v2 agent with sentiment analysis capabilities

## Key Questions & Answers

### ElizaOS v2 Features & Installation
**Q: How do I get the new beta working?**  
**A:** Follow these steps: 1. `npm create eliza@beta` 2. select project 3. cd project 4. `npm i` 5. `npx @elizaos/cli start` (answered by blackflame)

**Q: Is v1.0.0 alpha stable yet or should I stick with 0.25.9?**  
**A:** Not everything is smooth with this version either. (answered by spirit)

**Q: Is there any document summarizing how Eliza V2 has addressed the issues of V1?**  
**A:** Almost everything changed - there's a GUI, Tauri app, CLI, in-browser editor, no need for character files, simplified API, etc. (answered by yikesawjeez)

**Q: How can I start the server to control agents via API?**  
**A:** You can create a server using DirectClient as shown in the example code (answered by Ayush)

### Knowledge Implementation
**Q: Where should the Knowledge folder be located in the project for it to be detected when launching an agent?**  
**A:** Try `characters/knowledge/another_folder` (answered by Nooters)

**Q: Is there any example on how to use PDFs for knowledge?**  
**A:** Folder2knowledge — knowledge2character. Takes a PDF, puts it in plain text, then puts it in character knowledge. (answered by Midas)

### Spartan/DegenAI Status
**Q: When will we be able to speak with Spartan again?**  
**A:** "I'm trying to get spartan chatting before the v2 official launch. But there maybe other reasons I'm not aware of where we'd want to coordinate with other teams on that." (answered by Odilitime)

**Q: Is there any ETA for V2 official launch?**  
**A:** Beta phase will last about 2 weeks. (answered by rhota)

**Q: Will there be any token utility?**  
**A:** There are plans around ai16z utility forming, degenai utility could be part of that, but right now the focus is on getting open source functionality working in v2 and rolling Spartan out to as many communities as possible. (answered by rhota)

### Miscellaneous
**Q: How to apply for the launchpad?**  
**A:** https://tally.so/r/3Ev6XX - can get there from elizaos.ai > partners > apply (answered by yikesawjeez)

**Q: How do I use ollama to run agent correctly? I got embedding 404 error with llama3.2 and DeepSeek on ollama.**  
**A:** Check these video guides for Ollama setup (answered by chris.troutner)

## Community Help & Collaboration

### Knowledge Implementation Assistance
- **Nooters** helped **King Dam** with RAG knowledge folder issues by suggesting placing knowledge files in `characters/knowledge/agent_knowledge/` and converting PDFs to .md files for better processing
- **blackflame** provided step-by-step instructions for installing and starting the beta version to multiple users
- **natrix5942** offered to collaborate with **JohnMark13** on a PR for Supabase RAG knowledge implementation and share progress on fixing the issues

### Documentation & Visual Assets
- **Patt** provided specific feedback on images generated by **jin** for documentation and offered to help with img2img edits
- **shaw** suggested using Flux (fal.ai/models/fal-ai/flux-pro/v1.1-ultra) to **jin** for better image generation
- **Ka_yari** provided positive reinforcement to **jin** on design direction for documentation visuals

### Technical Support
- **Ayush** helped **alohabro** with REST API implementation by providing example code for creating a DirectClient server
- **chris.troutner** shared two video guides with **TrustWarden** for setting up Ollama with GPU and open-webui
- **spirit** provided the correct CLI installation command to multiple users: `npm install -g @elizaos/cli@1.0.0-beta.7`

### Community Organization
- **ben** modified social media copy drafted by **Patt** for official channels
- **accelxr** outlined immediate needs for DAO partners including quality content amplification, representation, culture improvement, and structured feedback
- **vincentpaul** created a structured introduction format to map skills to contribution areas for DAO partners

## Action Items

### Technical
- Fix RAG knowledge implementation issue where agent doesn't provide responses based on knowledge folder (Mentioned by King Dam)
- Implement solution for configuring slash commands in client-discord (Mentioned by HarryRiddle.eth)
- Investigate embedding 404 error with llama3.2 and DeepSeek on Ollama (Mentioned by TrustWarden)
- Fix Supabase RAG knowledge implementation issues (Mentioned by JohnMark13)
- Resolve plugin-sql version compatibility issues (Mentioned by Multiple users)
- Fix UUID constraint issues in knowledge chunks (Mentioned by JohnMark13)
- Implement search_knowledge function for adapters (Mentioned by kaisermerkle)
- Fix better-sqlite3 error when starting agents (Mentioned by natrix5942)
- Fix bugs to make Spartan functional in v2 (Mentioned by rhota)
- Get Spartan chatting before the v2 official launch (Mentioned by Odilitime)
- Complete v2 beta phase (approximately 2 weeks) (Mentioned by rhota)
- Implement automatic summarization/aggregator for Telegram (Mentioned by jin)
- Develop v2 for the Espresso hackathon (Mentioned by yikesawjeez)
- Implement Negation Game with harmonica.chat (Mentioned by yikesawjeez)
- Explore Hyperlane SVM integration to connect Solana with Arbitrum Nitro using Espresso DA (Mentioned by yikesawjeez)
- Implement token2022 functionality on Solana as a simplified alternative to the L2 approach (Mentioned by DorianD)
- Develop system for registering agents with character JSONs and social profiles using public keys (Mentioned by DorianD)

### Documentation
- Improve navigation in docs with mermaid flow charts and better organization (Mentioned by jin)
- Update documentation for knowledge directory structure (Mentioned by Multiple users)
- Create clear migration guide from v0.25.x to v1.0.0 (Mentioned by Multiple users)
- Document how to run multiple agents in one runtime (Mentioned by natrix5942)
- Create visual assets for ElizaOS documentation pages (Mentioned by jin)
- Add SVG images to documentation using AI assistance (Mentioned by jin)
- Review documentation one more time before v2 release (Mentioned by jin)
- Improve public information channels about the project (Mentioned by Toni)
- Create clearer guidebook to onboard more developers for ElizaOS V2 (Mentioned by Zolo)
- Update documentation to clearly explain V2 improvements over V1 (Mentioned by Avanc)

### Feature
- Explore capability to trigger multiple actions with a single sentence (Mentioned by Liu#001)
- Support for running multiple Telegram bots with different tokens (Mentioned by natrix5942)
- Improve quality of knowledge feature which "seems to work 10% ok" (Mentioned by Sabochee)
- Implement sentiment analysis in Spartan v2 agent (Mentioned by rhota)
- Roll out Spartan to as many communities as possible (Mentioned by rhota)
- Develop token utility as part of ai16z utility plans (Mentioned by rhota)
- Create mintable NFTs of software box art as swag/support mechanism (Mentioned by jin)
- Move Red or a Red stand-in to Eliza so she can participate in Discord (Mentioned by yikesawjeez)
- Activate DegenSpartan AI as an agent in Discord to improve morale (Mentioned by vincentpaul)
- Create a structured feedback system for constructive criticism (Mentioned by accelxr)
- Develop a program/guidelines for volunteer community contributors (Mentioned by accelxr)
- Create mechanism for agents to send messages that can later be used as IP for training new LLMs (Mentioned by DorianD)
- Implement stake-weighted voting system to secure new Elizas/plugins and partner with ecosystem teams (Mentioned by yikesawjeez)
- Design slashing mechanism for community members who stake tokens on teams that fail to deliver (Mentioned by yikesawjeez)
- Implement tokenomics integration with plugin ecosystem - "stake to plugin developer" mechanism (Mentioned by yikesawjeez)
- Create system for agents to stake tokens and report plugin usage metrics (Mentioned by DorianD)