Build agents with
OpenCode
Use OpenCode, the privacy-first open-source AI code assistant. Self-host, use any LLM, and keep full control of your data.
Setup Steps
Install OpenCode
Install the open-source AI code assistant
npm install -g opencode-cli Configure your LLM
Use any LLM provider (OpenAI, Anthropic, local models)
opencode config --provider anthropic --api-key YOUR_KEY Install Squads CLI
Install the Squads CLI globally
npm install -g squads-cli Initialize agents
Create the agent structure
squads init Why OpenCode?
Privacy-first
Your code never leaves your infrastructure. Perfect for sensitive codebases.
opencode --local-only Use any LLM
Connect to OpenAI, Anthropic, Gemini, or self-hosted models.
opencode config --provider ollama Open source
Audit the code, contribute features, fork and customize as needed.
github.com/opencode-ai/opencode Works with squads
Run agents using OpenCode as the execution engine.
squads run research --engine opencode Supported LLM Providers
Use any provider or run models locally
Anthropic Claude models via API
OpenAI GPT-4, GPT-4o via API
Google Gemini models via API
Ollama Local models (Llama, Mistral)
LM Studio Self-hosted models
Custom Any OpenAI-compatible API
Ready to build?
Initialize your project and start building
npm install -g squads-cli && squads init