Configuration files for OpenCode using the Regolo AI provider.
| File | Description |
|---|---|
opencode.json |
Main OpenCode configuration with Regolo provider setup, MCP servers, and permissions |
oh-my-opencode.json |
Oh-My-OpenCode plugin configuration with agent model assignments and categories |
- qwen3.5-122b - Main reasoning model
- qwen3-coder-next - Fast coding model (240k context)
- mistral-small-4-119b - Balanced model
- minimax-m2.5 - Large context model (130k)
- gpt-oss-120b - Alternative reasoning model
- context7 - Official documentation lookup
- grep_app - Real-world code examples search
- websearch (Exa) - Web search capabilities
- sequentialthinking - Structured reasoning
- fetch - URL content retrieval
- deepwiki - Repository knowledge extraction
| Agent | Model | Purpose |
|---|---|---|
| sisyphus | minimax-m2.5 | Main orchestrator |
| oracle | qwen3.5-122b | Read-only consultation |
| librarian | qwen3.5-122b | External documentation |
| explore | qwen3-coder-next | Codebase exploration |
| prometheus | minimax-m2.5 | Planning |
| metis | mistral-small-4-119b | Pre-planning analysis |
| momus | qwen3-coder-next | Quality review |
-
Copy configuration files to your OpenCode config directory:
cp opencode.json ~/.config/opencode/ cp oh-my-opencode.json ~/.config/opencode/
-
Set your Regolo API key:
export REGOLOAI_API_KEY="your-api-key"
-
Restart OpenCode
We strongly recommend creating an AGENTS.md file in your project root to provide behavioral guidelines for AI coding assistants. This helps reduce common LLM mistakes and improves code quality.
You can use the Andrej Karpathy-style guidelines as a starting point.
- Consistency: Ensures AI follows your coding standards
- Quality: Reduces over-engineering and unnecessary changes
- Efficiency: AI asks clarifying questions instead of guessing
- Safety: Prevents unwanted refactoring of working code
- OpenCode installed
- Regolo AI account and API key
- oh-my-opencode plugin (auto-installed via config)
