Claude: reviews and analysis

TLDR: Claude is Anthropic’s AI assistant for writing, reviewing documents and thinking a problem through with someone on the other side. There is a permanent free tier and the Pro plan starts at $17 a month billed annually. Its strength is long-form text where nuance matters, plus sustained work inside a project that holds your documents and instructions. Its weak spot is usage limits, which run out without ever telling you how much you had left.

What Claude is and how it works

Claude is Anthropic’s conversational assistant, from the same company behind Claude Code. What you get here isn’t an agent that touches your repository, it’s a chat window where you write, upload documents, ask for analysis and work the text until it’s usable. This is the generalist product, the one used by people who don’t code.

Its home turf is the document. Drafting a multi-page report, reviewing a contract for what’s missing, condensing fifty pages of regulation without dropping the caveats, rewriting someone else’s copy while keeping their voice. On that kind of job, where the output is judged by nuance rather than speed, is where paying users say it pulls ahead.

The feature that changes how you work is projects. A project is a workspace with its own chat history, its own behaviour instructions and a knowledge base where you upload the documents Claude should always have in front of it. Free accounts can create up to five. On top sits memory, which keeps context separated per project instead of mixing everything, with the option to review and edit what it kept and an incognito mode for conversations you don’t want stored.

It runs on web, iOS, Android, Mac and Windows, with connectors that pull data from other tools into the conversation. In February 2026 it reached number one on the US App Store’s free downloads chart, a fact that says little about quality and a lot about it no longer being a technical niche product.

What it’s like to use day to day

The use that pays off isn’t asking and copying. It’s pasting a draft, asking it to take the draft apart, arguing over two versions and keeping the third. That back and forth is the real workflow of everyone who pays for it.

The learning curve is about unlearning search-box habits. Type a three-line brief and expect a report, and you get correct, flat prose. Give it the context, the audience and the tone, and you get something publishable after one editing pass. Projects exist precisely so you don’t retype that context every morning, and setting the first one up properly is half an hour that pays back within a week.

The real friction is the quota. Every plan has a ceiling that refills on rolling windows, and the official help explains what drains it, long conversations, heavy files, connected tools and higher reasoning effort, without publishing per-plan numbers. In practice you run dry mid-task with no warning you could have read, and that’s the most repeated complaint across forums and app store reviews. The habit that stretches a quota furthest is opening a fresh chat when the topic changes instead of dragging a hundred-message thread along.

The other two frictions come up less often. Moderation turns strict on subjects any adult would handle fine, something visible in App Store reviews alongside complaints about suspended accounts with no support reply. And every version change spawns threads asking whether quality dropped, with more perception than measurement behind it.

Pricing and plans

There’s a permanent free tier and it covers small-scale work, with chat, web search, file uploads and up to five projects on a much shorter quota. Pro costs from $17 a month billed annually, $20 if you pay monthly, and it raises usage, adds long-running research and bundles Claude Code. Above that, Max plans start at $100 and multiply the quota for people who exhaust it daily.

The maths works in two steps. If your job is writing and analysis, $20 is the going rate in this category and the competition here is on text quality rather than price. If you also code, Pro includes the coding agent, and that bundle is what tips the decision against paying two separate subscriptions. What you can’t work out in advance is how much usage you’ll actually get, so squeeze the free plan for a few weeks before committing to the annual bill.

What happens to your conversations

The free tier can help improve Claude, but accepting that use is not an unavoidable condition for keeping it. On personal Free, Pro and Max accounts, Anthropic lets you choose whether chats help improve Claude. The switch sits under Settings, Privacy, “Help Improve Claude.” Turning it off stops new chats and coding sessions from being used in future model training. Conversations flagged by safety systems may still be used for safety and abuse detection.

The setting cannot remove data from training that has already started. Client documents or regulated information are better handled under a commercial product’s specific terms than by relying only on a consumer-account control.

Who it’s for (and who it isn’t)

Claude is for people who live off text: consultants writing reports, lawyers reviewing contracts, analysts digesting dense documentation, teachers preparing material, product people writing specs. The longer and more delicate the document, the clearer the gap. And if you write and code, Pro earns its price by including Claude Code at no extra cost.

It isn’t for anyone after photos, illustrations or generative video. Claude can create SVGs and diagrams inside Artifacts, but photorealistic visual generation still calls for a dedicated tool. Nor is it for anyone needing high, constant volume at a low price, because the Pro ceiling arrives sooner than expected and the jump to Max multiplies the bill fivefold. And if what you want is to compare answers from several models on one fee, YesChat covers that specific intent better than any single subscription.

Alternatives to Claude

The inevitable comparison is ChatGPT, which wins on everything built around the chat, image, voice and an app ecosystem, while Claude holds up better on long-form text when nuance counts. Gemini is the natural pick for anyone working inside Google Workspace who wants the assistant where their documents already live. And YesChat answers the case of using several models on a single fee. The full rundown is in Claude alternatives, and the category ranking in the ai assistants.

Frequently asked questions

Does Claude have a free version?

Yes, and it’s permanent. It covers chat, web search, file uploads and up to five projects, on a shorter quota than the paid plans. There’s no day-limited trial because the free tier already plays that role.

How is it different from Claude Code?

Claude is the chat assistant for writing, analysing and organising work. Claude Code is the agent that works on your repository from the terminal and the IDE. They share models and subscription, so a Pro plan gets you both.

Why do I run out of usage sooner than expected?

Because consumption isn’t counted in messages. Long conversations, heavy files and connected tools burn through it far faster, and the official docs acknowledge this without giving per-plan figures. Starting a new chat when the topic changes is the single adjustment that stretches a quota most.

Does it remember what I told it last week?

Inside a project, yes. Memory keeps context separated per project rather than mixing everything, you can review and edit what it kept, and incognito mode leaves out anything you don’t want stored.

Does it generate images or video?

It does not generate photos, illustrations or video. It can create SVGs, diagrams and other editable visuals inside Artifacts, an important distinction when you need a chart or interface rather than a photorealistic image.