Use LibreChat with llmrelay

Setup

1. 1 Copy librechat.example.yaml to librechat.yaml next to your docker-compose.yml.
2. 2 Under endpoints.custom, add an entry with name: 'LLM Relay' and provider: 'anthropic'.
3. 3 Set baseURL to https://api.llmrelay.dev and apiKey to '${LLMRELAY_API_KEY}'.
4. 4 List models explicitly under models.default and set models.fetch to false, since auto-fetch uses the OpenAI /models convention.
5. 5 Put LLMRELAY_API_KEY=*** in your .env file and restart the stack (docker compose up -d).
6. 6 Select LLM Relay in the endpoint dropdown in the UI.

Config

# librechat.yaml
endpoints:
  custom:
    # provider: 'anthropic' uses the native Anthropic client (/v1/messages)
    - name: 'LLM Relay'
      provider: 'anthropic'
      apiKey: '${LLMRELAY_API_KEY}'
      baseURL: 'https://api.llmrelay.dev'
      headers:
        anthropic-version: '2023-06-01'
      models:
        default:
          - 'claude-opus-5'
          - 'claude-sonnet-5'
        fetch: false
      titleConvo: true
      titleModel: 'claude-sonnet-5'
      modelDisplayLabel: 'LLM Relay (Claude)'

# .env
# LLMRELAY_API_KEY=***

Smoke test

Copy, paste, run. If you get JSON back with a message, you're good.

Run `docker compose up -d` then check the container logs for the custom endpoint being registered, and send one message on the LLM Relay endpoint in the UI.

Gotchas

• !With provider: 'anthropic' the baseURL must be the API root that the Anthropic SDK appends /v1/messages to, so do not include /v1 yourself.
• !Model auto-fetch does not work for the anthropic provider; you must list models explicitly and set fetch: false.
• !Leaving out provider: 'anthropic' falls back to the OpenAI-compatible client, in which case baseURL should be https://api.llmrelay.dev/v1 instead.

Related integrations