CAMI Documentation
Your AI Agent Guild Headquarters
Create specialized Claude Code agents from requirements, organize them in a clean workspace, and share them across your team. CAMI handles everything through natural language.
You: "I'm building a meditation app. Help me set up the right agents"
CAMI: Let me analyze your requirements...
I recommend: frontend, backend, mobile-native, ux, accessibility-expert
You: "Create those agents for my project"
CAMI: ✓ Created project at ~/projects/meditation-app
✓ Deployed 5 specialized agents
✓ Generated CLAUDE.md with project vision
Ready to build!
The Problem: Three Phases of Agent Chaos
Every developer building with Claude Code faces the same progression
"I don't know how to build agents"
Starting with Claude Code but unsure how to structure agents. What makes a good agent vs bad agent? How do I create my first specialized agent?
"My agents are scattered everywhere"
Created a few agents manually. Copy-pasting between projects. No version control or organization. Which version is the latest?
"My team needs to collaborate"
Multiple people creating agents. Need a standard library for the team. Want to track versions and updates across projects.
Quick Start
Get CAMI running in minutes with Go and make:
Install CAMI
Clone repository, run make install - creates ~/cami-workspace/ and binary
Deploy Your First Agent
Deploy agents to .claude/agents/ with automatic manifest tracking
Deployment guide →