Continue with a free LLM API

What Continue needs

Continue reads a YAML assistant file and treats an OpenAI-compatible endpoint as a first-class provider, with the key resolved from its secret store.

1. Each entry under models: needs provider: openai, an apiBase that serves /chat/completions, and the provider’s exact model id.
2. apiKey uses the ${{ secrets.NAME }} form, so the file can be committed while the value stays outside it.
3. roles decides where the model is offered. A small free model is usually fine for chat and unreliable for apply, and listing a role it cannot handle shows up as broken edits rather than an API error.
4. A model id that begins with @, as Cloudflare’s do, has to be quoted, because YAML reserves that character.

The file this produces

Running npx free-llm-api setup continue writes config.yaml. It is generated from the catalog, so the base URL and model id are the ones the provider documents rather than ones you have to look up. Here it is filled in for GroqCloud:

name: GroqCloud Free LLM
version: 0.1.0
schema: v1
models:
  - name: GroqCloud llama-3.3-70b-versatile
    provider: openai
    model: llama-3.3-70b-versatile
    apiBase: https://api.groq.com/openai/v1
    apiKey: ${{ secrets.GROQ_API_KEY }}
    roles:
      - chat
      - edit
      - apply

Nothing in that file is a key. The value is read from GROQ_API_KEY at run time, which is the one habit that keeps a credential out of a repository by construction.

Which free tiers fit Continue

22 providers here serve the OpenAI protocol and take no credit card, which makes them the ones worth trying first with Continue. The full list, including the ones that do ask for a card, is in the catalog.

What to expect on a free tier

Continue keeps a persistent context of open files, so the request size grows with the session rather than with what you typed. Token-per-day limits bind before request-per-day limits do.

Before blaming the client, check the key itself: the browser checker separates a rejected credential from a rate limit from an endpoint that is simply down, and each of those has a different fix.

Sources

• Continue configuration reference

Provider figures shown here come from each provider’s own documentation, reviewed on 2026-07-25. The methodology covers what is recorded and what is deliberately left blank.