# elizaOS Development Discord - 2025-04-28

## Overall Discussion Highlights

### Agent Configuration & Setup
- Users are actively setting up Eliza agents with various plugins, particularly Twitter integration
- The community is working through configuration challenges, especially around authentication and credentials
- ElizaOS is now TypeScript-based (v2.1.0), moving away from earlier Python implementations
- Configuration is split between technical settings (config.yaml) and personality/response settings (character files)

### Advanced Agent Capabilities
- Interest in agent-to-agent communication (swarms) and orchestration between multiple agents
- Discussion about MCP (Mission Control Protocol) integration with the Fleek plugin
- Hyperfy confirmed to be using the Fleek plugin for MCP client functionality

### Storage & Infrastructure
- Technical discussion about storage system architecture with 3 hot and 2 cold replicas for warm storage
- Debate about data replication approaches versus erasure coding for efficiency
- Implementation of Proof of Data Possession (PDP) for verification of distributed data chunks

## Key Questions & Answers

**Q: Is there a simple guide for setting up an example agent on Discord?**  
A: jin recommended using the full LLM context guide at https://eliza.how/llms-full.txt

**Q: What is the simplest way to get a Twitter agent working?**  
A: sam-developer provided GitHub link to social media manager code and basic setup instructions

**Q: Can we use Twitter agent without Twitter API?**  
A: sam-developer confirmed it's possible using cookies for authentication

**Q: Is it possible for agents to talk to each other with an orchestrator agent?**  
A: DeFine explained this is called a swarm, suggesting modifications to the org package or creating an Eliza project from CLI

**Q: What MCP client plugin is Hyperfy using?**  
A: jin and peezy confirmed they're using the Fleek plugin

**Q: How is the warm storage system designed for redundancy?**  
A: The plan includes 3 hot and 2 cold replicas, with hot copies chunked and distributed across many storacha nodes using Proof of Data Possession

**Q: Is ElizaOS Python or TypeScript based?**  
A: Ruby clarified that ElizaOS is now TypeScript-based (v2.1.0), and Python examples were from an early prototype

## Community Help & Collaboration

- **Twitter Agent Setup**: sam-developer helped Ryan-MoRtRed with Twitter agent configuration, providing GitHub links and explaining authentication options
- **Storage System Design**: Ruby provided technical feedback to inthiseconomy about their storage architecture, suggesting erasure coding instead of full replication
- **Twitter Mention Listener**: Ruby assisted Pill Bullman with configuring the Twitter plugin to monitor and respond to mentions, providing specific code for config.yaml
- **Documentation Clarification**: 0xbbjoker helped standard understand prompt formatting, explaining why logs appear flattened while actual formatting is preserved
- **MCP Plugin Identification**: jin and peezy helped Agent Joshua identify which plugin Hyperfy uses for MCP client functionality

## Action Items

### Technical
- Improve Twitter agent authentication process (Ryan-MoRtRed)
- Address Twitter account suspension issue after posting (Ryan-MoRtRed)
- Implement agent-to-agent communication framework (AD)
- Consider implementing erasure coding instead of full replication for better storage efficiency (Ruby)
- Address potential hotspot issues in shard management (Ruby)
- Complete stress testing on Filecoin Calibration network (inthiseconomy)
- Configure Twitter plugin with watchUsers and replyToMentions settings in config.yaml (Ruby)

### Documentation
- Create beginner-friendly guide for setting up example agents with plugins (BigBoss)
- Develop clearer instructions for Twitter agent setup (Ryan-MoRtRed)
- Improve MCP server tools documentation (Scooter)
- Create more accessible onboarding materials beyond video tutorials (BigBoss)
- Update outdated guides that reference Python implementation (Ruby)

### Feature
- Implement agent-to-agent communication framework (AD)