Auto Claude: Open-Source AI Project Management Platform
📖 Overview
Auto Claude transforms the Claude API into a fully autonomous virtual software agency using Kanban boards and parallel processing. Instead of babysitting a single AI chat, you become the manager of multiple AI agents working in parallel.
GitHub Repository: https://github.com/AndyMik90/Auto-Claude Discord Community: https://discord.gg/QhRnz9m5HE
🎥 Video Sources
| Video | Creator | Duration | Focus |
|---|---|---|---|
| Auto Claude Overview | AICodeKing | 8 min | Philosophy & quick tour |
| Auto Claude Deep Dive | André Mikalsen | 19 min | Detailed walkthrough & installation |
🎯 Core Philosophy
The Problem: Being a bottleneck when babysitting AI - you need to scale yourself, not just the AI.
The Solution: Shift from “chatbot” to “virtual software agency”
- You define what needs to be done
- AI handles how to do it
- Human reviews and merges
🔑 Key Features
| Feature | Description |
|---|---|
| Kanban Board | Tasks flow: Planning → In Progress → AI Review → Human Review → Done |
| Git Worktrees | Multiple agents work on different branches without merge conflicts |
| Agent Terminals | Command up to 12 “junior developer” agents simultaneously |
| Graph Memory + Semantic RAG | Indexes project, understands file relationships, builds knowledge graph |
| AI Self-Review | AI critiques its own work before human review |
| API Key Rotation | Multiple keys for handling rate limits |
| Changelog Generator | Automated changelogs from git history |
| Ideation Feature | AI-suggested features and automatic ticket generation |
📋 Combined Curriculum
AICodeKing Overview (8 min)
| Timestamp | Topic |
|---|---|
| 0:00 | Introduction - Ralph Wigum plugin recap |
| 0:45 | The Problem - Being a bottleneck |
| 1:15 | Introducing Auto Claude |
| 1:45 | Core Goals - Context amnesia, multi-threading |
| 2:30 | Installation - Electron app |
| 2:45 | Interface Overview - Kanban workflow |
| 3:15 | Task Creation |
| 3:45 | Git Worktrees for parallelization |
| 4:30 | Conflict Resolution |
| 5:00 | Agent Terminals (up to 12) |
| 5:45 | Graph Memory System |
| 6:30 | Ideation & Roadmap |
| 7:00 | Changelog Generator |
| 7:15 | API Key Rotation |
André Mikalsen Deep Dive (19 min)
| Timestamp | Topic |
|---|---|
| 00:00 | Introduction to Auto Claude |
| 00:39 | Getting Started |
| 01:04 | Using the Kanban Board |
| 02:11 | Task Management and AI Review |
| 06:09 | Agent Terminals and Session Management |
| 07:36 | Insights and Roadmap Functionality |
| 08:37 | Context and Memory System |
| 11:13 | Change Log and GitHub Integration |
| 12:49 | Advanced Settings and Integrations |
| 16:04 | Installing Auto Claude |
| 18:59 | Conclusion and Community |
🏗️ Architecture
Workflow Pipeline
Define Tasks → Kanban Board → AI Executes → AI Self-Review → Human Review → Merge
Git Worktrees Parallelization
- Each agent works in isolated worktree
- No merge conflicts during parallel execution
- Dedicated AI layer for conflict resolution when merging
Graph Memory System
- Indexes entire project
- Understands file relationships
- Builds knowledge graph of codebase
- Gets smarter the more you use it
💻 Installation
- Type: Electron app (cross-platform)
- Platforms: Windows, Mac, Linux
- Web option: Mentioned as available
🎯 Learning Objectives
- Understand the philosophy shift from chatbot to virtual agency
- How Kanban boards organize AI tasks
- How Git Worktrees enable parallel agent execution
- The Graph Memory and Semantic RAG system
- How to use Agent Terminals (up to 12 agents)
- The self-review workflow
- API key rotation for rate limits
- Ideation feature for project analysis
- Installation process
✅ Actionable Points
- Install Auto Claude (Electron app)
- Set up multiple API keys for rotation (especially for Claude Opus 4.5)
- Create Kanban board for project tasks
- Explore parallel agents via Git Worktrees
- Use Ideation feature for feature suggestions
- Join Discord community for support
- Configure GitHub integration
🔗 Related Topics
- Git Worktrees tutorial
- Claude API multi-session management
- Semantic RAG for code understanding
- AI coding agents comparison
- Kanban for AI workflows
🔗 Related Notes
- [[2025-12-22-aicodeking-opus-gemini-pro-coding-workflow]] - Opus + Gemini workflow (complementary model-switching strategy)
⭐ Rating & Review
After exploring:
- Quality (1-5): _/5
- Relevance (1-5): _/5
- Would recommend: Yes / No
- Best for: Developers wanting to scale AI coding workflows, those frustrated with single-threaded AI interactions
Captured: 2026-01-05 Sources: AICodeKing, André Mikalsen Repository: https://github.com/AndyMik90/Auto-Claude