GetRouterDocs
  • English

    • Continue

    Continue supports the OpenAI provider with a custom apiBase. This guide keeps the model on the OpenAI Chat Completions protocol.

    Integration status
    • Protocol: OpenAI Chat Completions (forced by configuration)
    • Configuration checked: July 15, 2026
    • Runtime verification: Pending

    1. Install Continue

    Install Continue from the VS Code or JetBrains plugin marketplace. Confirm whether your current version uses a local YAML configuration or a Continue Hub configuration.

    2. Configure the model

    Choose a MODEL_ID that supports OpenAI Chat Completions, then add the model to models in the current YAML configuration:

    name: GetRouter
version: 1.0.0
schema: v1

models:
  - name: GetRouter Model
    provider: openai
    model: MODEL_ID
    apiBase: https://getrouter.ai/v1
    apiKey: YOUR_API_KEY
    useResponsesApi: false
    roles:
      - chat
      - edit
      - apply

    Keep apiBase as https://getrouter.ai/v1; it is the API base URL, not a complete operation endpoint. For this model entry, useResponsesApi: false prevents Continue from automatically selecting the Responses API and fixes the protocol to Chat Completions.

    Use Continue's supported secret or environment-variable mechanism instead of committing a real key to repository configuration.

    If you intentionally switch to the Responses API, update the configuration to enable useResponsesApi and handle requests, responses, and streaming events according to the Responses protocol rather than Chat Completions.

    3. Validate the configuration

    Reload the IDE and select GetRouter Model:

    1. Test Chat.
    2. Test an explanation of selected code.
    3. Test a small Edit.
    4. Review the usage record.
    5. Test Agent mode separately only if the selected model supports tool calling.

    Multiple-model guidance

    Continue can assign different models to Chat, Edit, Apply, and Autocomplete. Enable only roles you have validated, and do not assume that one model is suitable for every role.

    Troubleshooting

    • Configuration does not load: Check YAML indentation, the schema version, and the active configuration source.
    • 401: Check whether the IDE process can read the configured secret or environment variable.
    • 404: Keep apiBase at https://getrouter.ai/v1; do not append a complete operation endpoint.
    • Protocol mismatch: Confirm that useResponsesApi: false is present and that MODEL_ID supports Chat Completions.
    • Autocomplete does not work: Autocomplete may require a different role, API, or model capability.