Your AI Terminal Coding Agent
AutoDevOS runs an autonomous agentic loop — read, edit, execute, and iterate with your codebase using natural language. Supports every major LLM.
Capabilities
Everything Your Workflow Needs
Built for engineers who want AI to do the heavy lifting — autonomously and securely
Agentic Loop
Autonomous multi-step loop — plan, execute tools, observe results, and iterate without manual intervention up to 100 turns.
Multi-Provider LLMs
OpenAI, Anthropic Claude, Google Gemini, Ollama, OpenRouter, LM Studio, vLLM, and any OpenAI-compatible endpoint.
11 Built-in Tools
Shell execution, file read/write/edit, grep, glob, web search, web fetch, memory, and TODO tracking — ready to use.
MCP Integration
Extend capabilities with any Model Context Protocol (MCP) server via stdio or HTTP/SSE transport.
Smart Context Management
Automatic conversation compression when approaching the model context window preserves task continuity.
Safety Controls
Approval policies, blocked destructive commands, path validation, and secret-variable filtering built in.
Lifecycle Hooks
Trigger shell scripts at before_agent, after_agent, before_tool, after_tool, and on_error events.
Session Persistence
Save and resume conversations with full context using /save and /clear in-session commands.
Loop Detection
Detects repeated identical tool calls and injects a circuit-breaker to prevent runaway loops.
Supported LLM Providers
Workflow
Up and Running in Minutes
Install & Authenticate
Install via pip or pipx, then run the interactive auth wizard to connect your preferred LLM provider and store your API key securely.
Open Your Project
Navigate to any codebase. The agent reads your project context, AGENTS.md instructions, and project-level config automatically.
Describe & Execute
Describe a task in plain English. The agent reads your code, makes edits, runs commands, and verifies results autonomously.
Built-in Toolkit
Tools Ready Out of the Box
The agent picks the right tool for every step automatically — no configuration needed
In Action
Watch the Agent Work
Community
Loved by Developers
AutoDevOS completely changed how I develop. It's like having a senior engineer always available.
Alex Chen
Full Stack Developer
The agentic loop is incredible — I give it a task and come back to it done. No hand-holding needed.
Sarah Williams
Software Engineer
Finally an AI terminal tool that lets me use local Ollama models for privacy-sensitive work.
Mike Johnson
DevOps Engineer
Ready to Ship Faster?
Install AutoDevOS and let the agent handle the repetitive work so you can focus on what matters.