Courses, curricula and practice platforms that cost nothing and are taught well. Everything here is genuinely free to learn from, with no course you have to buy to finish.
Thousands of hours of curriculum with certifications and real projects, all free with no upsell. Best for absolute beginners who need structure rather than a pile of links.
A full-stack path that makes you set up your own environment and build projects from a specification instead of following along. Harder than freeCodeCamp, and it produces people who can work unsupervised.
The Harvard introduction to computer science, and the best-produced course on this list by some distance. It teaches you how computers work, not just how to write JavaScript.
Actual MIT course materials, including lectures, problem sets and solutions, published openly for decades. No support and no certificate, so bring your own discipline.
The place to repair a shaky mathematics foundation before it blocks you in machine learning or graphics. Aimed at school level and useful at any age.
University of Helsinki course on modern JavaScript, React, Node and GraphQL, kept genuinely current. Free credits are available if you complete the exercises.
Practice exercises in more than sixty languages with free human mentoring on your submissions. The feedback is what makes it different from every other practice site.
Interactive lessons in a browser sandbox, which lowers the barrier for a first hour of code. The free tier is thin now, and much of the good material sits behind Pro.
Community-maintained skill trees for frontend, backend, DevOps and more. Not a course, a map, and it is the fastest cure for not knowing what to learn next.
The reference for HTML, CSS and JavaScript, plus learning paths that are better than most paid courses. If a tutorial contradicts MDN, trust MDN.
MIT lectures on the shell, git, editors and debugging, the tooling every course assumes you already know. A weekend here saves months of confusion.
Deep learning taught top down, with a working model in the first lesson and the theory afterwards. Ideal for programmers; frustrating if you want the mathematics first.
Free courses on transformers, diffusion models, agents and reinforcement learning, taught with the libraries the industry actually uses. Assumes you can already write Python.
Short data science micro-courses in hosted notebooks with free GPU time attached. Good for filling a specific gap in an afternoon rather than learning a field.
A complete computer science degree curriculum assembled from free courses, in a sensible order. Years of work with no deadlines, so it suits self-directed learners only.