OpenCode¶

Uses your configured OpenCode provider and model for inference.

Setup¶

1. Install OpenCode and ensure it's running (opencode in a terminal)
2. Sidekick connects to the local server on port 4096
3. Set sidekick.inferenceProvider to opencode in settings

How It Works¶

• Uses @opencode-ai/sdk to connect to the local OpenCode server
• Model selection is handled by your OpenCode configuration
• Tier values (fast/balanced/powerful) are passed as hints, but OpenCode's settings take precedence

Session Monitoring¶

OpenCode sessions are monitored from ~/.local/share/opencode/. Set sidekick.sessionProvider to opencode or leave as auto.

Troubleshooting¶

Connection issues¶

• Ensure OpenCode is running (opencode in a terminal)
• Sidekick connects to http://127.0.0.1:4096 by default