Continue

Continue is an open-source AI coding assistant that supports chat, autocomplete, and inline edits in VS Code and JetBrains.

Setup

Install Continue from the VS Code or JetBrains plugin marketplace
Open the Continue config: run Continue: Open config.yaml from the command palette (Cmd+Shift+P)
Add your ModelBridge models:

config.yaml

models:
  - name: Claude Sonnet 4-6
    provider: openai
    model: claude-sonnet-4-6
    apiBase: https://api.modelbridge.dev/v1
    apiKey: mb_live_your_key_here

  - name: GPT-5.4
    provider: openai
    model: gpt-5.4
    apiBase: https://api.modelbridge.dev/v1
    apiKey: mb_live_your_key_here

  - name: Claude Opus 4-6
    provider: openai
    model: claude-opus-4-6
    apiBase: https://api.modelbridge.dev/v1
    apiKey: mb_live_your_key_here

Save the file. Continue picks up changes automatically.
Open Continue chat, select your model, and test it.

Key points

Use provider: openai for all models routed through ModelBridge, even Claude models. ModelBridge speaks the OpenAI format.
Each model gets its own config. Unlike Cursor, Continue lets you set different apiBase and apiKey per model.
Set apiBase to end with /v1. Continue appends /chat/completions automatically.
If Continue tries to use the /responses endpoint for GPT models, add useResponsesApi: false to the model config.

Autocomplete

Configure a separate model for autocomplete (tab completion):

Autocomplete config

tabAutocompleteModel:
  provider: openai
  model: claude-sonnet-4-6
  apiBase: https://api.modelbridge.dev/v1
  apiKey: mb_live_your_key_here

Multiple roles

You can assign different models for different tasks:

Multiple roles

models:
  - name: Claude Sonnet 4-6
    provider: openai
    model: claude-sonnet-4-6
    apiBase: https://api.modelbridge.dev/v1
    apiKey: mb_live_your_key_here
    roles:
      - chat
      - edit

  - name: GPT-5.3 Codex
    provider: openai
    model: gpt-5.3-codex
    apiBase: https://api.modelbridge.dev/v1
    apiKey: mb_live_your_key_here
    roles:
      - apply