Playground
Every concept from the course can be validated against your own model here. Test whether it connects, how long the first token takes, how streaming chunks arrive, and what happens when the connection drops.
LLM Connection Lab
Fill in your own endpoint address and API key, and test the connection directly from your browser.Your API key stays in your browser only—our server neither receives, logs, nor forwards it—this page involves no server-side processing, requests go directly from your browser to the target service.
A few things worth trying
- Click "Test connection" first to verify your key and endpoint. Check the "handshake latency"—that's the time from request to response header.
- Then click "Send" and compare handshake latency to "first token latency". First-token latency is usually much longer because the model needs to think before speaking.
- Drag temperature to 0 and send the same prompt twice—the replies should nearly match. Then set it to 1.5 and try again to feel the randomness.
- Start a generation and click "Simulate network loss" halfway through. Watch how the incomplete reply stops. This is what the D1 section on "reconnection" is solving.
Don't have a key? OpenRouter gives you a free credit after signup—enough to run through these steps.