Skip to content

Saturday, October 10, 2026

Live
Loading the latest AI news…
AI Broadsheet Labs

Path 5 of 6 · People who can write a little code

Build with AI APIs

Call a model from your own code: OpenAI, Gemini and Claude side by side, then Microsoft's open course to put it together. Some Python helps.

Time
32 h
Steps
5 + 4 optional
Level
Intermediate
In French
1 of 9
Price
Free

You'll be able to

  • Call a model from code with an API key
  • Stream, structure and check what comes back
  • Add tools and document search (RAG)
  • Keep costs down and measure quality
5 steps
  1. Step 1CourseOpenAI

    OpenAI developer quickstart (opens in a new tab)

    Get an API key, install the library and make a first request; then send images and files and add tools such as web search and function calling.

    • about 30 min
    • Intermediate
    • Reading · Code
    • Free
    • English only

    Free to read; running the code calls an API that bills per use.

    Read it at OpenAI
  2. Step 2CourseGoogle

    Gemini API: getting started (opens in a new tab)

    Create a key and send a first request in Python, JavaScript or plain HTTP; then streaming, images, structured output and tools.

    Has a free tier, handy for practice.

    • about 30 min
    • Intermediate
    • Reading · Code
    • Free
    • English only

    Free tier with rate limits; a paid tier lifts them.

    Read it at Google
  3. Step 3CourseAnthropic

    Building with the Claude API (opens in a new tab)

    A full developer course in 13 modules: first requests, prompting, tool use, retrieval, MCP and agent workflows, with exercises. Needs Python and some JSON.

    • about 8 h
    • Intermediate
    • Video · Code
    • Free
    • English only
    • Provider certificate

    Free to read; running the code calls an API that bills per use.

Go further

Optional; not needed for the certificate

  1. Step 6CourseAnthropicOptional

    Anthropic courses on GitHub (opens in a new tab)

    Five notebook courses meant to be taken in order: API fundamentals, prompting, real-world prompting, prompt evaluations and tool use.

    • about 10 h
    • Intermediate
    • Code
    • Free
    • English only

    Free to read; running the code calls an API that bills per use.

  2. Step 7CourseOpenAIOptional

    OpenAI Cookbook (opens in a new tab)

    Open-source, runnable examples for common jobs (retrieval, structured output, evaluation, agents), mostly in Python.

    • about 2 h
    • Intermediate
    • Code
    • Free
    • English only

    Free to read; running the code calls an API that bills per use.

    Read it at OpenAI
  3. Step 8CourseHugging FaceOptional

    LLM Course (opens in a new tab)

    Hugging Face's open course on how language models work and how to fine-tune, evaluate and share them. Needs good Python.

    Go under the hood: how the models themselves work.

    • about 60 h
    • Advanced
    • Reading · Video · Code
    • Free
    • FR