Continue

The open-source AI code assistant you configure your way

Connect with the TokenMix CLI

The official TokenMix CLI generates a ready-to-paste config in one command — no manual environment variables:

npx tokenmix continue

Copy the output and paste it into the Continue extension settings. Prefer to configure it by hand? The full steps are below.

Introduction

Continue is a flexible, open-source AI coding assistant that works inside VS Code and JetBrains. It provides inline tab-completions, a chat sidebar, and the ability to edit multiple files simultaneously. Unlike cloud-locked tools, Continue stores its configuration in a plain YAML file you control. Connect it to TokenMix to use any model you want.

Prerequisites

• VS Code 1.84+ or a JetBrains IDE (IntelliJ, PyCharm, WebStorm, etc.)
• A TokenMix account with an API key starting with sk-tm- → Register at tokenmix.ai → Dashboard → API Keys

Setup Steps

1. Install the Continue extension

VS Code: Open the Extensions panel (Cmd+Shift+X / Ctrl+Shift+X), search for "Continue", and click Install.

JetBrains: Go to Settings → Plugins → Marketplace, search for "Continue", and click Install.

2. Open the configuration file

Click the Continue icon in the Activity Bar to open the sidebar. Click the gear icon (⚙) → "Open config.yaml".

This opens ~/.continue/config.yaml in your editor.

3. Add a TokenMix model

Edit ~/.continue/config.yaml to add a model under the models key:

models:
  - name: Claude Sonnet via TokenMix
    provider: openai
    apiBase: https://api.tokenmix.ai/v1
    apiKey: sk-tm-your-key-here
    model: claude-sonnet-4-6
    roles:
      - chat
      - edit

To add a second model for tab completions:

  - name: DeepSeek Coder via TokenMix
    provider: openai
    apiBase: https://api.tokenmix.ai/v1
    apiKey: sk-tm-your-key-here
    model: deepseek-v4-pro
    roles:
      - autocomplete

4. Save and reload

Save config.yaml. Continue auto-reloads within a few seconds. You should see your new model names appear in the Continue chat sidebar model selector.

Verify

In the Continue chat sidebar, select your new model and type:

Explain what this file does in 3 bullet points.

Continue should respond using the TokenMix-powered model. For completions, open any code file and press Tab after a partial statement.

FAQ

Q: Why provider: openai for Claude models? Continue uses provider: openai to signal the OpenAI-compatible API format. TokenMix exposes a unified OpenAI-compatible endpoint for all models including Claude, so this setting works correctly.

Q: Can I use different models for chat vs. autocomplete? Yes. Set roles: [chat, edit] on your primary model and roles: [autocomplete] on a faster/cheaper model (e.g., claude-haiku-4.5 or deepseek-v4-pro).

Q: My API key appears in plain text in config.yaml — is that safe? The file lives in your home directory and is only readable by your user account. For extra safety, you can store the key in an environment variable and reference it as apiKey: $TOKENMIX_API_KEY.