HTTP
Every AI service speaks HTTP. You need to understand methods, headers, status codes, and bodies to build and debug integrations.
HTTP is ordering at a restaurant — you place an order (request), the kitchen works on it, and brings food or an error back (response).
Visual Workflows
Start here — study each diagram, then use + / − to zoom if needed.
Overview
Scroll inside the frame · use + / − to zoom
Scroll inside the frame · use + / − to zoom
Key Takeaways
- 1.HTTP is the request-response protocol behind every web API — including every LLM provider you'll ever integrate with.
- 2.A client sends a method (GET/POST), URL, headers, and optional body.
- 3.The server returns a status code and response body.
- 4.LLM APIs use POST with JSON.
Real Example
Scenario
Your chat app returns 401 Unauthorized when a user sends a message.
What you would do
The Authorization header with the Bearer token was missing. Adding it returns 200 with the LLM response.
Practice Task
Send a GET request with curl -v https://httpbin.org/get. Then send a POST with curl -X POST -H "Content-Type: application/json" -d '{"name":"test"}' https://httpbin.org/post. Write down the status code and response body for each.
Commands
Commands to Remember
curl -X GET https://api.example.com/data # send a GET request to read datacurl -X POST -H "Content-Type: application/json" -d '{"key":"val"}' URL # send POST with JSON bodycurl -v URL # verbose mode — see headers, status code, and response
Cheat Sheet
Quick recap
quick ref- •Methods: GET read · POST action · PUT replace · PATCH update · DELETE remove
- •Every request has: URL, Headers, Body (optional)
- •Key headers: Authorization Bearer token, Content-Type application/json
- •Status codes: 200 OK · 401 Unauthorized · 429 Rate Limited · 500 Server Error
- •Stateless — each request is independent
- •LLM APIs use POST + JSON · always set timeouts
- •HTTPS = HTTP + TLS encryption
Common Mistakes
- ✕No timeout on HTTP calls
- ✕API keys in URL instead of headers
- ✕Ignoring 429 rate limits