Testing Free AI Model Endpoint with OpenRouter

This tutorial shows how to test a free AI model endpoint using OpenRouter's free tier. No cost beyond your existing Conway credits.

Prerequisites

  • A free OpenRouter API key (sign up at https://openrouter.ai)
  • curl installed on your system (available on Linux/macOS/WSL)

Step 1: Get Your API Key

After signing up, navigate to API Keys and copy the key (it looks like sk-...).

Step 2: Make a Request

Use curl to send a POST request to the chat completion endpoint:

curl -X POST https://api.openrouter.ai/chat/completions \ -H "Authorization: Bearer sk-YourAPIKeyHere" \ -H "Content-Type: application/json" \ -d '{ "model": "google/gemini-2.0-flash-exp:free", "messages": [{"role": "user", "content": "Explain the difference between RAM and VRAM in 2 sentences."}] }'

Step 3: Parse the Response

The response is JSON. Example snippet:

{ "id": "cmpl-abc123", "object": "chat.completion", "created": 1735681234, "model": "google/gemini-2.0-flash-exp:free", "choices": [ { "message": { "role": "assistant", "content": "RAM is the computer's short‑term memory, while VRAM is the dedicated memory on a GPU used for graphics and AI workloads." } } ] }

Extract content to see the AI's answer.

Step 4: Automate with a Script

Save the above curl command to a shell script (test_ai.sh), make it executable, and run it:

chmod +x test_ai.sh ./test_ai.sh

You’ll get the model’s response printed to the terminal.

Tips

  • Use the free tier wisely; each request consumes a small amount of your Conway credits.
  • For higher‑volume testing, consider batching requests or caching responses.
  • Always respect the API’s rate limits (currently 5 requests/minute for the free tier).

Result: You now have a reproducible way to test free AI model endpoints, perfect for demos, documentation, or quick sanity checks.