LibreChat

If you are accessing an existing LibreChat web instance deployed by your team or community, you can start chatting immediately:

Option A: User-Provided API Keys

If the LibreChat instance allows users to provide their own API keys:

1. Open and log in to your LibreChat site;
2. Click User Profile / Settings in the bottom left ➔ Account / API Keys;
3. Locate OpenRely or Custom Endpoints, and paste your API Key from the OpenRely console;
4. Return to the chat window to select a model and begin chatting.

Option B: Pre-configured Endpoint

If the instance administrator has already configured OpenRely globally:

1. In the chat window, click the Model Dropdown Menu at the top and select the OpenRely group;
2. Choose your target model (reasoning, coding, or lightweight general models);
3. Type your prompt to experience fast streaming responses and multimodal analysis!

If you manage your own LibreChat instance, simply declare the custom endpoint in your librechat.yaml:

version: 1.1.5
cache: true

endpoints:
  custom:
    - name: "OpenRely"
      apiKey: "${OPENRELY_API_KEY}" # or your raw sk-... key
      baseURL: "https://api.openrely.ai/v1"
      models:
        default: ["gpt", "claude", "gemini", "deepseek"]
        fetch: true # dynamically fetch accessible models from /v1/models
      titleConvo: true
      titleModel: "gpt"
      modelDisplayLabel: "OpenRely"
      dropParams: ["stop"]

[!NOTE] If using the ${OPENRELY_API_KEY} environment variable, simply add OPENRELY_API_KEY=sk-your_key_here to your .env file and restart the instance. No complex setup required!