Claude

Anthropic's assistant: for writing, analysis and building things, reading up to 1M tokens at once. A free mini-course from the basics to the power moves.

Whisperer's verdict: 4.5 / 5

What it nails

• Reads very long documents (1M tokens)
• Builds apps, charts and files you can use
• Asks before acting when you want it to

Watch out for

• No built-in image generation

Claude is the AI assistant made by Anthropic. It reads very long documents (up to 1M tokens), writes and edits, and builds working things: interactive charts, apps, spreadsheets and slide decks. In September 2026 Anthropic merged chat and Cowork into one Claude that "can now figure out what a task needs", so the same conversation can answer a question or produce a finished document.

Start here: the basics

1. Sign in at claude.ai or in the desktop and mobile apps. The Free plan includes web search, file creation, code execution and memory.
2. Write the way you would brief a colleague. Anthropic's golden rule: "Show your prompt to a colleague with minimal context on the task and ask them to follow it. If they'd be confused, Claude will be too."
3. Attach files. Up to 20 per chat and 500MB each: PDFs, Word, Excel, CSV, images and more.
4. Ask for the finished thing, not just advice: "make the spreadsheet", "build the chart".

Lesson 1: say how far to go

"Being specific about your desired output can help enhance results." Anthropic's own before and after:

Create an analytics dashboard

Example by Anthropic · source ↗ · Anthropic's own "less effective" example.

Create an analytics dashboard. Include as many relevant features and interactions as possible. Go beyond the basics to create a fully-featured implementation.

Example by Anthropic · source ↗ · Anthropic's "more effective" version.

Lesson 2: give the reason, not just the rule

"Providing context or motivation behind your instructions can help Claude better understand your goals."

NEVER use ellipses

Example by Anthropic · source ↗

Your response will be read aloud by a text-to-speech engine, so never use ellipses since the text-to-speech engine will not know how to pronounce them.

Example by Anthropic · source ↗

Which model and plan to pick

Power moves

1. Make files, not just answers

Claude can "execute code to create and work with files directly in your conversations": Excel with formulas, PowerPoint, Word, PDF and charts, on every plan.

Extract all the data from this PDF into an Excel spreadsheet and create a summary chart.

Example by Anthropic · source ↗ · An official sample prompt. Attach your PDF first.

2. Artifacts: things you can click and share

An Artifact is "anything Claude makes for you that you'd put in front of someone: a design, a deck, a document, a dashboard, or a small interactive tool." They start private and can be published with a link.

an interactive visualization showing which countries export the most bananas per capita

Example by Zapier · source ↗

3. Projects that remember your world

A Project holds your files and instructions for one goal. Since 17 September 2026, Claude "scopes the request, delegates the work, coordinates parallel threads, reviews the outputs, and assembles the finished result" inside a project.

4. Build real apps

With Claude Code, Claude plans, writes and tests software on its own. Ask for the finished product and set the rules up front.

i want you to build a fully researched and beautiful isochronic map that lets me pick various cities and see real isochronic lines based on real data. I want the design to be unique. You should take into account airports (and travel time to and from airports) trains, walking, driving. The data does not need to be live but should be real based on your research and data. You can start with a few cities but more general is better, this should be an entirely new project.

Example by One Useful Thing (Ethan Mollick) · source ↗ · Made with Claude Fable 5 in Claude Code.

Build this 3D game, for the browser.
This repo is configured to serve static files so make sure there is an index.html that loads everything else.
Make sure it is mobile-friendly (touch controls, works well on small screens).
You have an OpenAI API key and access to their image generation model APIs, use that for textures to use with your 3D models. Docs here: https://developers.openai.com/api/docs/guides/image-generation - use gpt-image-2
Work independently - do not ask me to make any further design decisions. Make sure the game is fun, a little surprising, has good raccoon heist vibes, and is visually pleasing.
Commit and push as often as possible so I can preview your work - start with an index.html that presents a title screen, then build from there.
Append to a notes.md file as you work, including your changes to that as part of every commit.

Example by Simon Willison's Weblog · source ↗ · Made with Claude Fable 5 in Claude Code on the web. He also attached screenshots describing the game.

5. Connect your apps and let it work

Connectors let Claude "access your apps and services, retrieve your data, and take actions within connected services". Research runs "multiple searches that build on each other" and returns cited answers (paid plans).

connect to my Gmail and help me prep for the MBA seminar I am giving on Monday the 21st, including building some presentation and demos as inspiration. Answer any outstanding messages on the topic.

Example by One Useful Thing (Ethan Mollick) · source ↗ · Claude's answer is the top one; below it is ChatGPT Work. Claude prepared drafts and asked before sending.

6. Structure big prompts

• Show examples. "Examples are one of the most reliable ways to steer Claude's output format, tone, and structure." Use 3 to 5, wrapped in <example> tags.
• Use XML tags such as <document> and <instructions> to separate parts of a long prompt.
• Put long documents first and your question at the end: Anthropic measured answers "up to 30%" better.
• Say what to do, not what not to do: "Your response should be composed of smoothly flowing prose paragraphs."
• Ask it to check itself: "Before you finish, verify your answer against [test criteria]".

Generate an SVG of a pelican riding a bicycle

Example by Simon Willison's Weblog · source ↗ · A famous informal test of drawing with code. Made with Claude Fable 5.1 at maximum reasoning.

FAQ

Sources: Claude models overview · Claude pricing · Cowork is now Claude · Prompting best practices · Artifacts help