ai-integrations-openrouter

You can access OpenRouter chat completions without an API key by using the Replit AI Integrations proxy. This proxy provisions managed LLM access, handling the authentication internally so you can query the unified endpoint directly.

The Replit AI Integrations proxy is a middleware layer that provides OpenRouter-compatible API access. It routes your chat completion requests to long-tail providers via a base URL without requiring you to provision your own API keys.

To set up the OpenRouter proxy integration, provision the integration in the JavaScript sandbox, copy the provided templates, install dependencies, and configure the required environment variables to wire the chat completion endpoint.

Yes, this integration is specifically designed for Agent Stack scenarios. It provides safe, managed LLM access and environment provisioning suitable for complex AI integration workflows and automations.

Yes, the proxy supports long-tail AI providers by routing requests through a unified OpenRouter endpoint. It uses a base URL and internal provisioning to connect you to various providers without needing individual API keys.

The main limitation is dependency on the Replit AI Integrations proxy for authentication and routing. You do not control the API keys directly, which means access is managed and provisioned by the proxy environment rather than your own account.