OpenCode is a fully open-source (MIT-licensed) AI coding agent built by the team behind SST. It connects to 75+ model providers — Claude, GPT, Gemini, DeepSeek, or local models via Ollama — so you control cost and model choice instead of being locked to one vendor. Standout features include LSP-aware diagnostics fed back to the model, background subagents, a research-focused Scout agent, and fully air-gapped deployment for regulated environments. With 160,000+ GitHub stars and 7.5 million monthly developers, it's the most-adopted open-source coding agent built to date.
Developers who want a free, open-source coding agent that works with any AI model, including local ones.
- Fully open-source (MIT-licensed) and connects to 75+ model providers — genuine control over cost and model choice
- LSP-aware diagnostics fed back to the model is a distinctive technical feature few competitors offer
- Fully air-gapped deployment available for regulated environments
- Massive, proven adoption — 160,000+ GitHub stars and 7.5 million monthly developers
- You still pay for whichever underlying model API you connect it to — the tool is free, but usage isn't necessarily zero-cost
- Terminal/IDE-first workflow has more setup than a fully polished commercial editor
Is OpenCode really free, including the AI models it uses?
The tool itself is free and open-source, but you bring your own API keys for whichever models you connect — Claude, GPT, Gemini, DeepSeek, or fully free local models via Ollama if you want zero marginal cost.
What is Scout in OpenCode?
A research-focused subagent built into OpenCode for gathering external information as part of a coding task, running in the background alongside the main agent.
