LibreChat custom endpoint for UberLLM (librechat.yaml)

Steps

1. Make a key on your keys page and add credit.
2. Add the key to the .env file in LibreChat's folder:

UBERLLM_API_KEY=ull_your_key

Use this name, not OPENAI_API_KEY, so the built-in OpenAI endpoint does not pick it up.

3. Add UberLLM to librechat.yaml (same folder as .env):

version: 1.3.18
cache: true
endpoints:
  custom:
    - name: "UberLLM"
      apiKey: "${UBERLLM_API_KEY}"
      baseURL: "https://api.uberllm.dev/v1"
      models:
        default: ["deepseek/deepseek-v4.1-flash", "google/gemini-3.5-flash", "anthropic/claude-sonnet-5.5"]
        fetch: true
      titleConvo: true
      titleModel: "current_model"
      modelDisplayLabel: "UberLLM"

If your file already has endpoints: custom:, add only the - name: "UberLLM" block to that list. Keep your file's own version line if it is newer.

4. Docker: make sure the file is mounted. In docker-compose.override.yml:

services:
  api:
    volumes:
      - type: bind
        source: ./librechat.yaml
        target: /app/librechat.yaml

5. Restart: docker compose down && docker compose up -d (or restart the npm process).
6. Pick UberLLM in the endpoint menu and choose a model.

What the fields do

• baseURL: LibreChat adds /chat/completions to it, so stop at /v1.
• models.default: the models always shown. fetch: true asks UberLLM for its whole list on top of those.
• titleModel: "current_model": chat titles are written by the model you are chatting with. Set a cheap id such as deepseek/deepseek-v4.1-flash instead to keep title costs near zero.
• Want each user to bring their own UberLLM key? Set apiKey: "user_provided" and LibreChat asks them for it.

Troubleshooting

• UberLLM does not appear in the menu: LibreChat did not read the file. Check it is next to .env (or at CONFIG_PATH), mounted in Docker, and read docker compose logs api for a config error. A bad file stops LibreChat from starting.
• 401: ${UBERLLM_API_KEY} is empty inside the container. Check the .env line and restart.
• 404 model not found: ids need the maker in front, for example deepseek/deepseek-v4.1-flash.
• 402 insufficient credits: add credit on the credits page.
• A model rejects a setting: drop it for this endpoint with dropParams: ["stop"] (or ["user"]).

Keep reading

• Reference page: LibreChat doc
• Open WebUI setup
• One API key for all LLMs