Auto Claude is an autonomous AI coding companion designed to accelerate software development by automating planning, coding, and validation tasks. It targets developers of all skill levels, from "vibe coders" to experienced engineers, offering a significant boost in output and code quality through its agent-driven workflow. The application provides autonomous task execution, isolated development environments, and self-validation capabilities, allowing users to focus on higher-level objectives while agents handle feature development and bug fixing.

How It Works

Auto Claude employs a multi-phase agent pipeline: Spec Creation, Implementation, and Merge. Before coding, agents analyze the project, gather requirements, and generate a detailed specification. The implementation phase involves coding agents executing the plan, followed by QA agents validating the work. A key architectural choice is the use of Git worktrees to isolate each task in a separate workspace, ensuring the main branch remains undisturbed. Additionally, a Memory Layer, powered by FalkorDB via Docker, utilizes a hybrid RAG system to retain and leverage codebase insights across sessions, enhancing agent decision-making.

Quick Start & Requirements

The recommended way to use Auto Claude is via its Desktop UI.

• Prerequisites: Node.js 18+, Python 3.10+, Docker Desktop, Claude Code CLI (npm install -g @anthropic-ai/claude-code), a Claude Pro or Max subscription, and a Git-initialized project repository.
• Installation: Follow the setup guide for Node.js, Python, and Docker. Clone the repository, set up the Python backend (uv venv && uv pip install -r requirements.txt), start the FalkorDB memory layer (docker-compose up -d falkordb), and then install and launch the UI (pnpm install, pnpm run build && pnpm run start).
• Links: Node.js, Python, Docker Desktop. Detailed setup guides are available in guides/DOCKER-SETUP.md and guides/CLI-USAGE.md.

Highlighted Details

• Parallel Agents: Run multiple autonomous coding tasks concurrently.
• Self-Validating Code: Integrated QA agents verify code quality and acceptance criteria before user review.
• Isolated Workspaces: Utilizes Git worktrees to safely manage parallel development without impacting the main branch.
• AI Merge Resolution: Automatically resolves merge conflicts using AI when integrating changes back into the main branch.
• Memory Layer: A hybrid RAG system using FalkorDB provides persistent, cross-session context for agents.

Maintenance & Community

Auto Claude has an active community on Discord for support and discussion. Contribution guidelines are provided in CONTRIBUTING.md. The project acknowledges inspiration from Anthropic's Autonomous Coding Agent.

Licensing & Compatibility

The project is licensed under the AGPL-3.0 (GNU Affero General Public License v3.0). This license requires attribution, makes source code of modifications and network-served derivatives available under AGPL-3.0, and prohibits closed-source usage without a separate commercial license.

Limitations & Caveats

A paid Claude subscription (Pro or Max) is mandatory for Claude Code access. Docker Desktop is a strict requirement for the Memory Layer. The AGPL-3.0 license imposes significant copyleft obligations, potentially restricting adoption in closed-source commercial environments without obtaining a separate license. Windows users may require additional C++ build tools for certain installations.