The best free way to learn Claude is Anthropic's own material. Start with the free Claude Academy courses (many of the same courses are on Skilljar at anthropic.skilljar.com), then use the Claude documentation as your reference, the GitHub tutorials and Claude Cookbook for hands-on practice, and the Claude blog and Anthropic news to keep up.
That covers almost everything you need, and none of it costs money to read. The only costs show up when an exercise asks you to call the Claude API, which is billed by usage. This guide walks through each free resource, what it covers, who it suits, and a simple order to take them in, whether you use Claude for everyday work or plan to build with it.
Where Are Anthropic's Free Claude Courses?
Anthropic's courses live in two places with largely overlapping catalogs.
- • Claude Academy at academy.claude.com. The home page describes itself as "Free courses on Claude and AI, from Anthropic." The old address, anthropic.com/learn, now redirects here. The Claude Code documentation also points to it as "free self-paced courses."
- • Anthropic Courses on Skilljar at anthropic.skilljar.com. Skilljar is a third-party online course platform, and Anthropic's catalog page there is labeled "powered by Skilljar." Its header links back to Anthropic Academy. On the Claude 101 course page, the sign-up button reads "Register | FREE," and the curriculum ends with a certificate of completion module.
If you searched for "Skilljar blog" or "Claude blog" and landed somewhere unexpected, this is the reason: Skilljar is the course host, not a blog. The Claude blog is a separate site, covered further down.
Which Claude Academy Course Should You Take First?
The catalog lists lesson counts and estimated hours for each course. Here are the ones most people should look at first, with the figures shown on the Claude Academy course list:
- • Claude 101 (13 lessons, about 2.5 hours). Using Claude for everyday work: starting conversations, getting better results, organizing work with projects, artifacts and skills, connecting your tools, and research. It needs only a Claude account on any plan, though the course page notes that some hands-on features, such as Cowork and Research, require a paid plan.
- • AI Fluency: Framework and foundations (14 lessons, about 4 hours). Not about Claude features at all, but about working with AI systems "effectively, efficiently, ethically, and safely." A good pick if you want judgment, not button locations.
- • AI capabilities and limitations (13 lessons, about 3.5 hours). An introduction to how AI works, useful context before you trust any model with real work.
- • AI Fluency for small businesses (8 lessons, about 3.5 hours). The AI Fluency material framed for business owners who want efficiency without losing sight of their mission and values.
- • Introduction to Claude Cowork (14 lessons, about 2.5 hours). Working with Claude on real files and projects, including plugins, skills and multi-step tasks.
There are also versions of AI Fluency for educators, students, nonprofits, builders and creative work, so check the full catalog if one of those fits you better.
What Are the Best Free Claude Courses for Developers?
If you plan to write code that calls Claude, or to use Claude as a coding assistant, these courses go deeper:
- • Claude Code 101 (12 lessons, about 1.5 hours). Using Claude Code in your daily development workflow.
- • Claude Code in action (9 lessons, about 1 hour). Running longer Claude Code sessions: steering, configuring, automating and verifying the work.
- • Claude Platform 101 (13 lessons, about 1.5 hours). Building on the Claude Platform from the basics up.
- • Building with the Claude API (67 lessons, about 9 hours). The big one. It covers API basics and structured output, prompt engineering and evaluation, tool use, retrieval-augmented generation, prompt caching, Model Context Protocol, and agent workflows. The course page lists Python proficiency, basic JSON handling and an Anthropic API key as prerequisites, so expect some API usage cost for the hands-on parts.
- • Introduction to Model Context Protocol (10 lessons, about 1 hour). Building MCP servers and clients in Python, covering tools, resources and prompts. If MCP is new to you, our explainer on what MCP is and why it matters gives the plain-English version first. A follow-up course, Model Context Protocol: Advanced Topics, is listed on Skilljar.
- • Introduction to agent skills (6 lessons, about 1 hour). Building, configuring and sharing reusable Skills in Claude Code.
- • Introduction to evals (4 lessons, about 1 hour). How to test whether your prompts and agents actually work.
Lesson counts and hours change as Anthropic updates courses, so treat these as the figures shown on the catalog at the time of writing, not fixed numbers.
What Can You Learn From the Claude Documentation?
The documentation is free, always current, and the most authoritative source on how Claude works. There are three main sets.
- • Claude Platform docs at platform.claude.com/docs. For developers. The home page offers a quickstart, model overviews, the API reference, SDKs, and guides to features such as tool use, vision, structured outputs, prompt caching and streaming, plus sections on evaluation, guardrails and cost. Older docs.claude.com links now redirect here.
- • Claude Code docs at code.claude.com/docs. Installation, a quickstart, and guides to memory files, skills, hooks, MCP, subagents and automation. The docs themselves are free to read, but note that most Claude Code surfaces require a Claude subscription or an Anthropic Console account to actually run.
- • Claude Help Center at support.claude.com. For everyday users. Articles are grouped by topic, including the Claude apps, plans and usage limits, Claude Code, connectors, the API and Console, and privacy.
If you only bookmark one developer page, make it the release notes, linked from the docs home as "What's new." Claude changes quickly, and that page is where new features are announced for builders.
What Is the Best Free Guide to Prompting Claude?
Prompting is the skill that pays off on every plan, free or paid. Anthropic publishes several free resources on it.
- • Prompting best practices. The current reference for prompting Claude's latest models. Anthropic's prompt engineering overview calls it "the living reference," covering clarity, examples, XML structuring, role prompting, thinking and prompt chaining, with model-specific notes.
- • Prompt engineering overview. A short page on when prompt engineering is the right fix and when it is not. It points out that latency and cost problems are sometimes easier to solve by choosing a different model.
- • Interactive prompt engineering tutorial. Nine chapters plus an appendix, moving from basic prompt structure and being clear and direct through using examples, avoiding hallucinations and building complex prompts. Each lesson ends with a playground for editing prompts. A lighter Google Sheets version exists too. Anthropic's docs note it was written for Claude 3 models, so follow the best practices page where the two differ.
- • Best practices for prompt engineering on the Claude blog. A readable article covering being explicit, giving context, using examples and allowing uncertainty, plus advanced methods like chain of thought and prompt chaining.
To practice the basics in our own words, try our hands-on tutorial on writing effective prompts, which walks you through building a reusable prompt template.
Where Can You Practice Building With Claude for Free?
Reading only gets you so far. These GitHub resources are free to clone and study.
- • Claude Cookbook. A collection of practical guides and code examples, also on GitHub as anthropics/claude-cookbooks under an MIT license. Recipes cover agent patterns, tool use, evals, retrieval, vision, skills, extended thinking and integrations.
- • Anthropic courses repository. Five notebook courses, in a suggested order: Anthropic API fundamentals, the prompt engineering interactive tutorial, real world prompting, prompt evaluations, and tool use. The repo warns that the exercises make API calls, so you pay for usage, and notes the courses often use Claude 3 Haiku to keep costs low.
- • Claude Quickstarts. MIT-licensed starter projects, such as a customer support agent, a financial data analyst and computer use demos. You need a Claude API key to run them.
The code is free. Running it against the Claude API is not, which is why the free courses and docs are the better starting point if your budget is zero.
Which Blogs Should You Follow to Keep Up With Claude?
These are the official sources worth following, all free:
- • The Claude blog. It describes itself as "News, announcements, and stories from Anthropic about Claude and the people building with it," with filters for product updates, best practices, perspectives, insights and enterprise. This is the best answer to "where is the Claude blog."
- • Anthropic News. Company announcements, model releases, policy, safety and research news.
- • Anthropic Engineering. Technical posts from the team that builds Claude's infrastructure, focused on building, evaluating and deploying AI agents. Their Building effective agents post is a good starting point for agent design.
- • Anthropic Research. Work from teams including alignment, interpretability, societal impacts and economics. Heavier reading, but it explains the thinking behind how Claude is built and tested.
Do You Need a Paid Claude Plan to Learn Claude?
No. Claude's pricing page lists a Free plan at $0, described as "Free for everyone," with limited usage that resets every 5 hours and every week. It includes chat on web, desktop and mobile, web search, file creation, code execution, memory and artifacts. That is enough to practice prompting and work through most of Claude 101.
You will hit limits in three places:
- • Some features taught in courses, such as Cowork and Research, need a paid plan.
- • Most Claude Code surfaces need a subscription or a Console account.
- • Any course or repo that calls the Claude API needs an API key and is billed by usage.
Not sure Claude is the right tool for your task in the first place? Our Which AI for the Job? picker helps you match a task to a model before you invest the study time.
A Simple Free Learning Path for Claude
Here is an order that works for most people:
- • Step 1, everyday use. Take Claude 101 on Claude Academy and practice on the Free plan. Read the Help Center articles for anything that confuses you.
- • Step 2, better prompts. Read Prompting best practices, then work through the interactive tutorial, the Google Sheets version if you do not code.
- • Step 3, judgment. Take AI Fluency: Framework and foundations, or the version for your role.
- • Step 4, if you build software. Add Claude Code 101 or Building with the Claude API, keep the Platform docs open as your reference, and use the Cookbook when you need a working example.
- • Ongoing. Follow the Claude blog and Anthropic News so you know when features change.
The Bottom Line
You do not need to pay for a course to learn Claude. Anthropic gives away its training through Claude Academy (also hosted on Skilljar), keeps its documentation free and current, publishes tutorials and recipes on GitHub, and explains new features on the Claude blog. Start with Claude 101 and the prompting guide, add the developer material only if you build, and keep the docs as your source of truth.
Sources
- • Claude Academy (Anthropic)
- • Anthropic Courses on Skilljar
- • Claude 101 course page on Skilljar
- • Building with the Claude API course page (Claude Academy)
- • Claude Platform documentation
- • Claude Code documentation overview
- • Claude Help Center
- • Anthropic, Prompt engineering overview
- • Anthropic, Prompting best practices
- • Anthropic, Interactive prompt engineering tutorial (GitHub)
- • Anthropic courses repository (GitHub)
- • Claude Cookbook
- • Claude Quickstarts (GitHub)
- • Claude blog, Best practices for prompt engineering
- • Claude pricing
Tell us what is costing you the most time. We will map out exactly what your business needs. Free, no obligation.
Build An Agent