AI Made Simple · New to the terms? Use our AI glossary.
Need a hand writing a letter, understanding a confusing explanation or making a plan? Claude is an AI assistant from Anthropic that you can talk to by typing ordinary sentences. You do not need to learn computer commands.
This guide takes you from opening the website to trying your first useful request. Start with something small, leave out private information and treat the answer as a draft to check.
What you need before you begin
You can use Claude in a web browser on your computer, tablet or phone, without installing an app. Official mobile and desktop apps are also available. You must be 18 or over and in a supported location. See Claude’s getting-started guide.
Have your email account and mobile phone handy. Anthropic currently requires new users to verify a phone number by text message; a landline will not work. The number must be from a supported location. These requirements are explained in its phone verification guidance.
Set up your account, step by step
- Open the official website. Type claude.ai into your browser’s address bar. For a first visit, use this direct address rather than an advert offering an AI download.
- Choose how to sign in. Select Continue with Google to use a Google account, or enter your email address and choose Continue with email.
- Check your email if you chose that option. Open the login link you requested, preferably on the same device. If you open it on another device, follow the instructions to enter the resulting code on the original device. Check your junk folder if nothing arrives.
- Complete the account prompts. Read the terms and privacy choices. For phone verification, enter the code sent by text and select Verify code. Keep login links and codes private.
Claude uses Google sign-in or emailed links rather than a separate Claude password. The official login instructions explain both routes. Screen layouts may differ between devices.
Check this privacy setting first
Open your name or account menu, then choose Settings and Privacy. Look for Help Improve our AI models. Switch it off if you do not want your chats used for future model training.
Anthropic says switching this off stops new chats being used for future training and stops previously stored chats being used in future training runs. It does not undo training already started or completed. Conversations flagged by safety systems may still be used for safety purposes. Read Anthropic’s explanation of this setting.
Whatever you choose, do not paste passwords, bank details, medical records or confidential documents into a practice chat. Replace real names and addresses with placeholders such as “[name]” and “[address]”. Turning off training is not a promise that nothing is stored.
Try your first useful prompt
A prompt simply means the question or instruction you give Claude. Type this into the chat box, then use the send button:
Help me draft a polite email asking a repair company when they can visit to look at a leaking tap. Keep it under 100 words and use plain British English. Leave placeholders for my name and any appointment details. Do not invent a booking or a price.
Read the result. Does it say what you meant? You can reply, “Make it shorter” or “Make it sound less formal.” These follow-up requests help you shape the draft. Check it before copying it into your email app.
For better answers, include three things: what you want, useful background and how you want the answer presented. You might request three ideas, a short checklist or one step at a time. You can use everyday language.
Other simple things to try
- Explain a term: “What does ‘cloud storage’ mean? Explain it without technical words and give one everyday example.”
- Make a checklist: “Help me list what to pack for a two-night visit to family. Group it into clothing, toiletries and other essentials.”
- Improve your writing: “Make this message clearer without changing what I mean.” Add a short, non-sensitive example.
If an answer is confusing, say which part confused you. There is no need to know the right technical term before asking.
Do you need to pay?
Start with the free plan. Paid individual plans, including Pro and Max, offer more usage and additional features, but paying does not remove every limit. Check the official plan comparison before subscribing.
Claude’s free plan has usage limits. The allowance can vary with demand, and Claude tells you when you reach a limit. You can wait rather than upgrade. There is no need to choose a paid plan just to practise the examples here.
Check answers and stay alert to scams
Claude can give a confident answer that is wrong, outdated or based on an invented quotation. Anthropic recommends checking important claims and reading original sources rather than relying on Claude alone. See its guidance on incorrect answers.
Check dates, figures, links and anything you plan to send. For medical, legal or money decisions, consult a suitable professional. Asking Claude “Are you sure?” is not an independent check.
Do not rely on Claude to certify that a message is genuine. Avoid links in suspicious emails, and visit an organisation’s official website yourself instead. The National Cyber Security Centre recommends this safer approach.
Your first goal is simple: get one useful draft or explanation, check it, and decide whether it helped. You can explore the other features later.
Sources and check date
Checked on against the official Claude, Anthropic and National Cyber Security Centre pages linked above. Account requirements, menus, privacy options and plan limits can change.