How to use Continue through one API

Continue is the open-source AI assistant that lives inside VS Code and JetBrains. It is great at autocomplete, chat, and inline edits — but the moment you want to compare a GPT answer against Claude, or drop in DeepSeek for cheap bulk work, you hit the usual wall: a separate account, billing page, and API key for every provider.

There is a simpler path. Because Continue speaks the OpenAI API format, you can point it at a single gateway and reach every major model through one base URL and one key.

One endpoint, every model

AnyModel is an OpenAI-compatible gateway. You configure it once with:

• base_url: https://anymodel.org/v1
• one API key that reaches GPT, Claude, Gemini, DeepSeek, GLM, Kimi, Qwen, and Grok

To switch models inside Continue you change the model id — nothing else. No new keys, no new providers, no config rewrites. Want to see which model fits a task and a budget? The model comparison page lays out the options side by side.

Configuring Continue

Continue does not have a one-line installer the way some terminal tools do, so you set it up manually — which takes about a minute. Continue reads its configuration from ~/.continue/config.yaml (older versions use config.json). Add AnyModel as an OpenAI-compatible provider:

models:
  - name: GPT (AnyModel)
    provider: openai
    model: gpt-5
    apiBase: https://anymodel.org/v1
    apiKey: YOUR_API_KEY
    roles: [chat, edit, apply]

  - name: Claude (AnyModel)
    provider: openai
    model: claude-sonnet-4-5
    apiBase: https://anymodel.org/v1
    apiKey: YOUR_API_KEY
    roles: [chat, edit, apply]

  - name: DeepSeek (AnyModel)
    provider: openai
    model: deepseek-chat
    apiBase: https://anymodel.org/v1
    apiKey: YOUR_API_KEY
    roles: [chat, edit, apply]

The key detail is provider: openai paired with apiBase. That tells Continue to use the standard OpenAI request shape and send it to AnyModel instead of OpenAI's servers. Reload Continue (or restart your editor) and all three models appear in the model dropdown — each backed by the same key.

Note: the one-line installer (bash <(curl -fsSL "https://anymodel.org/i?tool=codex") <YOUR_API_KEY>) only covers codex, claude, opencode, and hermes. For Continue, Cursor, Windsurf, Zed, Cline, Aider, and Gemini CLI, use the manual OpenAI-compatible setup above.

Verify it works

Before wiring it into the editor, you can confirm the key and endpoint with a quick request:

curl https://anymodel.org/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5","messages":[{"role":"user","content":"Say hi from Continue"}]}'

If you get a JSON reply, your Continue config will work too — it sends the exact same request under the hood.

Why route Continue this way

You get to use the right model for each job — a frontier model for tricky refactors, a cheaper one for boilerplate and commit messages — without juggling accounts.

Free to start, pay only for what you use

Signing up gives you 1,000,000 tokens free, and linking Telegram brings the total to 6,000,000 — no credit card required. After that it is pay-per-token: no subscription, no monthly minimums.

If you work with sensitive code, you can enable Ghost Mode, an opt-in setting that issues zero-retention API keys. With Ghost Mode on, your prompts and responses are not stored on our side — only a token counter runs. (To be clear: the underlying model provider still receives your prompt, so this is not absolute privacy — but nothing is retained by us.)

Want more setup guides for other editors and CLIs? Browse the blog.

Get started

Configure Continue once, then switch between GPT, Claude, Gemini, and the rest by changing a single line. Create your free account and start coding with every model behind one API.