Skip to main content
OpenCode is an open-source terminal-based AI coding agent built by the SST team. The opencode runner invokes opencode run as a subprocess and supports over 75 AI providers — including GitHub Copilot and local models via Ollama — with no API key required to get started.

​
Requirements

  • OpenCode CLI installed (npm install -g opencode-ai)
Install OpenCode:
Verify the installation:

​
No API key required

OpenCode works out of the box with:
  • GitHub Copilot — if you have an active Copilot subscription, opencode will use it automatically
  • Local models via Ollama — run models entirely offline with no external API
  • OpenCode Zen — opencode’s own hosted model access: pay-as-you-go, 37+ models, no provider key needed. Get your key at opencode.ai/auth
If you want to use a specific cloud provider (Anthropic, OpenAI, etc.), you can optionally supply an API key. But it is not required to start.

​
Setup

Or choose opencode when prompted by the interactive wizard. During remoteagent init, the wizard will ask:
  • API key — optional; choose from:
    • Skip — not needed if you are using GitHub Copilot, a local model, or another pre-configured provider
    • OpenCode Zen — pay-as-you-go access to 37+ models, no separate provider key needed. Get your key at opencode.ai/auth
    • Anthropic / OpenAI / Google — paste your provider API key
  • Default model (e.g. anthropic/claude-sonnet-4-5) — optional; leave blank to use opencode’s own default
  • Small model for lightweight tasks — optional
  • Multiagent mode — yes/no

​
How it works

RemoteAgent.CHAT injects configuration at runtime using the OPENCODE_CONFIG_CONTENT environment variable — opencode’s native config injection mechanism. RemoteAgent.CHAT does not write any config files to disk. The runner executes:
If you chose OpenCode Zen during remoteagent init, the provider block is injected automatically and the default model is set to opencode/claude-sonnet-4-6:
Only the settings collected during the wizard are injected. Any additional opencode configuration you set in your own opencode.json file will be picked up by opencode automatically — RemoteAgent.CHAT does not interfere with it. Output from stdout is forwarded to Telegram as it arrives.

​
OpenCode’s own configuration

opencode has a rich native configuration system (opencode.json). Any settings not exposed in the RemoteAgent.CHAT wizard — custom keybindings, themes, provider-specific options, agent behaviors — can be configured directly in that file. opencode reads it automatically on every run. See OpenCode’s documentation for the full list of supported configuration options and providers.

​
Pros and cons

​
Updating OpenCode

Restart the agent after updating: