Continue

Configure Continue’s local YAML, apiBase, model roles, and explicit Chat Completions behavior.

Install, configure, and test

  1. Install Continue from VS Code Extensions (official instructions: https://docs.continue.dev/ide-extensions/install), then open its sidebar. In the local configuration selector, open the configuration file using its edit/gear control. The user-level file is ~/.continue/config.yaml on macOS/Linux or %USERPROFILE%\.continue\config.yaml on Windows. Edit YAML, not the deprecated config.json format.

  2. Create an API key in the routerra.ai console (https://routerra.ai/keys) and copy a model ID available to it. Merge the example below into your existing models list. Replace YOUR_API_KEY and YOUR_MODEL_ID. provider: openai selects the compatible adapter; apiBase is https://routerra.ai/v1, without /chat/completions. apiKey supplies gateway authentication; it is not a Continue account credential.

  3. Save the file and select the local configuration and its routerra model in Continue. roles limits this entry to chat, edit, and apply. Autocomplete and embeddings need separately configured models with the corresponding roles and supported endpoints. Add capabilities: [tool_use] only for a model and route that actually support tools; Agent mode requires that capability. Do not enable image_input without vision support.

  4. Start a new chat and ask a short project question. Reload the IDE window if changes are not reflected, then inspect the gateway request log. useResponsesApi: false keeps this entry on Chat Completions, including model families for which Continue otherwise selects Responses automatically.

If the model is missing, check YAML indentation and the active configuration. If requests unexpectedly use /responses, verify the setting is inside the selected model entry. Successful chat does not prove that an embedding or autocomplete role is configured.

name: routerra
version: 1.0.0
schema: v1
models:
  - name: routerra
    provider: openai
    model: YOUR_MODEL_ID
    apiKey: YOUR_API_KEY
    apiBase: https://routerra.ai/v1
    useResponsesApi: false
    roles:
      - chat
      - edit
      - apply