Les bases des API
-
Build a Study Assistant Without Overspending
Build a personal study assistant with a chat API, grounded answers, useful quizzes, and a budget you can control.
-
Call Several AI Providers Through One Endpoint
Connect once, test models from different providers, and track each request without turning a small project into a large bill.
-
Send Images to Vision Models, Step by Step
Learn how to prepare images, build a vision API request, check the response, and control spending through MyRelay.
-
Debug Request Logs Without Leaking Secrets
A practical walkthrough for logging AI requests safely, tracing failures, and spotting retries that quietly increase your spending.
-
Concurrency Limits: A Practical Walkthrough
Understand why concurrency limits exist, then build a simple queue, handle retries, and keep parallel AI requests within your budget.
-
Retry Flaky AI Requests Without Wasting Your Budget
A step-by-step walkthrough for handling network failures with bounded retries, sensible timeouts, backoff, and spending checks.
-
Count Tokens Before You Send a Request
A practical walkthrough for estimating input tokens, reserving output room, and checking request costs before spending your API budget.
-
Function Calling Through a Relay Endpoint
Build a small function calling loop through MyRelay, validate tool requests, return results, and keep token costs predictable.
-
Use the OpenAI Node.js SDK with a Custom Endpoint
Connect the OpenAI Node.js SDK to MyRelay, verify your first request, and keep retries, tokens, and testing costs under control.
-
Keep API Keys Out of Your Git History
A practical walkthrough for storing API keys safely, checking commits, and handling leaks without putting your development budget at risk.
-
Streaming AI Responses with Server-Sent Events
Build an SSE response flow step by step, handle interruptions, and track token costs without making your first streaming project expensive.
-
Handle Rate Limits Without Wasting Your Budget
A step-by-step walkthrough for pacing AI requests, retrying safely, saving progress, and keeping script costs predictable.
-
Your First API Call with cURL, Explained
Make your first AI API request with cURL, understand what each part does, and check the result without wasting your budget.