aifromzero.dev
All 15 modules live

smolagents Mastery — Code-First AI Agents with smolagents: From Your First CodeAgent to Production

Build a real code-first agent, module by module, and learn every moving part of smolagents — from the ReAct loop to secure sandboxing, multi-agent teams, and production telemetry.

15 modules Intermediate Python Free, no paywall No GPU required
Start with Module 1 Companion labs ↗

About this course

This is a free, 15-module course for Python developers who use LLM APIs but have never built an agent. Across the series you build Quill, a code-first data analyst — it writes and runs Python in a sandbox to clean, analyze and visualize data, browses the web for context, cross-checks its findings, and returns a cited report with charts — while covering the whole surface of Hugging Face smolagents: code-as-action, secure code execution, tools and MCP, multi-agent teams, vision, agentic RAG, deployment, and observability.

No certification gate here — this is the definitive, code-first smolagents course. Every lab lives in the companion repo, pinned to smolagents 1.26.0, and the full cumulative test suite runs on a laptop with no GPU. Start free with a Hugging Face token, or swap in any provider or a local model.

Syllabus

  1. 01 What Are Code Agents? From LLM Calls to smolagents (smolagents, Module 1) 12 min read Read
  2. 02 Your First CodeAgent: The ReAct Loop, Step by Step (smolagents, Module 2) 13 min read Read
  3. 03 Tools: Giving Quill New Powers (smolagents, Module 3) 13 min read Read
  4. 04 Models and Providers: What Powers Your Agent (smolagents, Module 4) 13 min read Read
  5. 05 Running Untrusted Code Safely: Sandboxing smolagents (smolagents, Module 5) 14 min read Read
  6. 06 Give Your Agent a Memory: State, Steps, and Inspecting Runs (smolagents, Module 6) 13 min read Read
  7. 07 Make Your Agent Reliable and Frugal: Planning and Building Good Agents (smolagents, Module 7) 13 min read Read
  8. 08 Reliable Agents: Structured Output, Validation, and Errors (smolagents, Module 8) 13 min read Read
  9. 09 Plug Quill Into Any Ecosystem: MCP, the Hub, and Tool Interop (smolagents, Module 9) 14 min read Read
  10. 10 Multi-Agent Systems: Quill Gets a Research Team (smolagents, Module 10) 13 min read Read
  11. 11 Vision and Multimodal: Quill Reads Charts and the Web (smolagents, Module 11) 13 min read Read
  12. 12 Agentic RAG: Grounding Quill in a Knowledge Base (smolagents, Module 12) 13 min read Read
  13. 13 Deploying Quill: Gradio UI, the Hub, and the CLI (smolagents, Module 13) 13 min read Read
  14. 14 Observability and Evaluation: Is Quill Any Good? (smolagents, Module 14) 16 min read Read
  15. 15 Capstone: Ship Quill, a Production-Grade Code Agent (smolagents, Module 15) 16 min read Read