Configures OpenCode to work with the OmniRoute AI gateway service, enabling access to multiple AI models from various providers including Qwen, Codex, GitHub and Kiro.
When to use
Use this skill when you want to:
- Connect OpenCode to the OmniRoute AI gateway
- Access multiple AI models through a single interface
- Configure all available OmniRoute models in OpenCode
- Switch between different AI providers easily
Prerequisites
- OpenCode version 1.2.0 or higher
- Valid OmniRoute API key
- Network access to OmniRoute service at http://138.2.29.190:20128
Instructions
- Obtain your OmniRoute API key
- Run the skill to initiate configuration
- Enter your API key when prompted
- Confirm the OmniRoute service endpoint (default: http://138.2.29.190:20128)
- Allow the script to configure OpenCode settings
- Verify configuration by testing a model
Configuration Details
The skill will:
- Create/update ~/.config/opencode/opencode.json with OmniRoute provider
- Add all available models (Qwen, Codex, GitHub, Kiro) to the configuration
- Update authentication credentials in ~/.local/share/opencode/auth.json
- Set up default model omniroute/qw/qwen3-coder-plus
Available Models
After successful configuration, you can use models such as:
- omniroute/kr/claude-sonnet-4.5
- omniroute/kr/claude-haiku-4.5
- omniroute/qw/qwen3-coder-plus
- omniroute/qw/qwen3-coder-flash
- omniroute/cx/gpt-5.2-codex
- omniroute/gh/gpt-4o
- And over 100 other models available through OmniRoute
Verification
The skill will automatically test the configuration with a simple API call to verify the connection works properly.