OpenOrca

OpenOrca is an autonomous AI coding agent that runs in your terminal. It connects to local LLM servers (LM Studio, Ollama, or any OpenAI-compatible API) and uses 39 built-in tools to read, write, and execute code — not just describe what to do, but actually do it.

Think of it as a local, private, open-source alternative to cloud-based AI coding assistants.

Key Capabilities

• Autonomous agent loop — the LLM plans, acts, observes results, and iterates up to 25 turns per request
• 39 built-in tools — file I/O, shell execution, git operations, web search, GitHub integration, and more
• Works with any local model — Mistral, Llama, DeepSeek, Qwen, or any model served via OpenAI-compatible API
• Native + text-based tool calling — auto-detects whether your model supports function calling and falls back gracefully
• Streaming with live thinking — see tokens arrive in real-time, or collapse thinking with Ctrl+O
• Plan mode — have the model plan before executing, then approve/modify/discard
• Session management — auto-save and restore conversations
• Context management — auto-compaction when approaching context limits
• Project instructions — drop an ORCA.md file in your project root for persistent instructions
• Custom slash commands — create .orca/commands/*.md files to define project-specific commands
• File checkpoints — automatic snapshots before file edits, with /checkpoint to list, diff, and restore
• Auto memory — session learnings are saved and loaded into future sessions automatically
• Permission glob patterns — fine-grained allow/deny rules like Bash(git *) and Write(src/**)
• Hooks — run custom shell commands before/after tool execution
• Permission system — approve tool calls by risk level
• Sub-agent spawning — delegate focused tasks to independent agent instances

Supported Platforms

Platform	File
Windows x64	openorca-win-x64.zip
Linux x64	openorca-linux-x64.tar.gz
macOS x64	openorca-osx-x64.tar.gz
macOS Apple Silicon	openorca-osx-arm64.tar.gz

Where to Start

I want to...	Go to...
Install and run OpenOrca for the first time	Installation > Quick Start
Understand which models work	Model Setup
Learn all commands and tools	Commands & Shortcuts > Tool Reference
Configure OpenOrca	Configuration
Troubleshoot a problem	Troubleshooting
Contribute or extend OpenOrca	Architecture > Adding Tools > Contributing