Claude Code official course (Anthropic)
For: Engineers using Claude Code as their primary AI coding assistant
Free as in actually free — not "free trial," not "free to audit but the certificate costs $49." The 29 picks below are the ones we keep recommending to friends starting from zero.
The best free AI courses in 2026 cluster around three sources: Anthropic and OpenAI's own vendor-published tutorials (interest-aligned with the vendor — useful but single-vendor), DeepLearning.AI's short courses (mostly free, broad coverage, taught by the actual framework authors), and Coursera's "audit" mode (free access to content, paid only for certificates). Outside these, the field is mostly free-trial-then-paywall — useful for the trial, less useful as durable resources.
Pick by where you are: zero-to-fluent on prompting starts with Anthropic's tutorial. Build-an-agent starts with the LangGraph short course. LLM internals starts with the Coursera Generative-AI-with-LLMs audit.
For: Engineers using Claude Code as their primary AI coding assistant
For: Developers building production AI agents in 2026
For: Engineers seriously committing to AI agent engineering as a craft
For: Engineers committed to LangGraph for production agents
For: Engineers who have used the OpenAI API but never built an agent loop
For: Engineers planning to fine-tune or self-host LLMs
For: Engineers who plan to fine-tune or self-host LLMs
For: Developers writing their first LLM-powered feature
For: Engineers who plan to train, fine-tune, or research LLMs at depth
For: Engineers without classical ML background entering AI work
For: Non-technical founders and PMs starting from zero
For: Engineers building production MCP servers for internal tools
For: Engineers whose basic RAG works in dev but fails in prod
For: Engineers prototyping multi-agent workflows
For: Anyone whose primary LLM is Claude (or who builds with Anthropic API)
For: Engineers about to commit to a vector DB choice
For: Engineers shipping their first LLM-with-tools app
For: Engineers whose basic RAG works in dev but fails in prod
For: Engineers building structured LLM apps but not yet full agents
For: Engineers on the OpenAI stack
For: Engineers wanting the conceptual map of agentic design patterns
For: Career-switchers entering AI engineering with formal credential value
For: Engineers building self-hosted RAG (privacy/cost-sensitive)
For: Engineers who learn by doing, not by deriving
For: Engineers learning vector DBs and RAG from first principles
For: Working developers wanting to learn the 2026-grade AI coding workflow
For: Engineers at AWS-stack companies
For: Engineers and PMs at GCP-stack companies
For: Non-engineers at Azure-heavy enterprises needing a cert