Use Continue with RunBridge AI

Connect Continue to RunBridge AI using its OpenAI-compatible provider and Chat Completions endpoint.

Prerequisites

Install the official Continue extension for VS Code.
Create an active API key in the RunBridge AI dashboard.
Choose an exact model ID that supports Chat Completions and streaming. Agent mode also requires tool calling.

Store your key locally

Open or create ~/.continue/.env (Windows: %USERPROFILE%\.continue\.env) in a local editor. Add your key there:

RUNBRIDGE_API_KEY=your-runbridge-api-key

Replace the placeholder only in this private file. Restrict file access to your user and keep it outside the project repository. Continue resolves the secret reference in the next step from its supported local secret sources; see the official secrets guidance.

Configure the model

Open Continue’s configuration selector, choose Local Config, and use the gear icon to open config.yaml. The default location is ~/.continue/config.yaml (Windows: %USERPROFILE%\.continue\config.yaml). Merge this model into an existing models list, or use the complete configuration below for a new local config:

name: RunBridge AI
version: 1.0.0
schema: v1
models:
  - name: RunBridge AI Chat
    provider: openai
    model: your-model-id
    apiBase: https://api.runbridge.ai/v1
    apiKey: ${{ secrets.RUNBRIDGE_API_KEY }}
    useResponsesApi: false
    roles:
      - chat

Replace your-model-id with the model you chose. The provider identifier stays openai; RunBridge AI Chat is your display name. useResponsesApi: false keeps this configuration on /v1/chat/completions, including model families that Continue otherwise sends to Responses. Do not append /chat/completions to apiBase.

Enable Agent mode

After confirming the model supports tool calling, add the following fields to that model entry:

    capabilities:
      - tool_use

Continue uses this capability to offer tool use for custom model IDs. Add image_input only if your selected model also accepts image input. Do not use a chat-only model for autocomplete or embeddings; those require separately configured models with the appropriate roles.

Verify the connection

Save the configuration and select RunBridge AI Chat in Continue. In Chat mode, send: Reply with “RunBridge connected”. Check the request in your RunBridge AI usage records. If you enabled Agent mode, test a small tool-assisted task in a disposable project and review any proposed edits before applying them.

Troubleshooting

Symptom	Check
Secret not found	Check the spelling of `RUNBRIDGE_API_KEY`, the `.env` location, and reload the editor after updating secrets. A workspace `.env` can take precedence over the global one.
YAML error or model missing	Check indentation, keep one `models` list, and select the local configuration you edited.
401	Verify the secret’s value and the key’s access in the RunBridge AI dashboard.
404 or unexpected Responses request	Keep `apiBase` at `/v1` and `useResponsesApi: false` inside the model entry.
Agent mode unavailable	Confirm real tool-calling support before adding `capabilities: [tool_use]`. A configuration flag cannot add support to a model.

See Continue’s official OpenAI provider, local configuration, and YAML reference.

Get started

Coding agents

Agent platforms

Editor extensions

Use Continue with RunBridge AI

Prerequisites

Store your key locally

Configure the model

Enable Agent mode

Verify the connection

Troubleshooting

​Prerequisites

​Store your key locally

​Configure the model

​Enable Agent mode

​Verify the connection

​Troubleshooting

Prerequisites

Store your key locally

Configure the model

Enable Agent mode

Verify the connection

Troubleshooting