marcella

In plain words · technology · unofficial

How Claude works

Claude in plain words: what this AI is, how Anthropic trains it and what happens between your question and its answer. Nine steps, an unofficial walkthrough based on Anthropic’s public materials. The diagram plays by itself; drag the scale below it.

In short

Claude is a language model made by Anthropic. It was trained on a large body of text to predict the next word and then further trained to be helpful and honest. When it answers, it reads the whole conversation as context, can think before replying and can search the web when needed.

Marcella, web and bot developer · updated 17 November 2026

Step 01 / 09

Pretraining

First Anthropic trains the model on a huge amount of text and code: public web pages, data from third parties and labelers, chats of users who allowed training on them, and data the company creates itself. At this stage there is one task — guess the next piece of text.

Technically: this gives a base model. It already knows the language and many facts but cannot yet hold a conversation like an assistant.

Glossary · tap a word and the diagram will show it

In short

Claude is an AI assistant and a family of language models from the American company Anthropic. The company was founded in 2021 by former OpenAI employees; the first version of Claude was shown in March 2023. Claude writes and edits text, answers questions, works through documents and images, writes program code. It runs at claude.ai, in desktop and mobile apps and through the API.

Inside Claude is a large language model, as in other chatbots: it writes the answer one token at a time, and its knowledge and skills live in its weights. First the model is trained to continue text on a huge dataset, then fine-tuned to be a helpful and safe assistant. Anthropic calls its approach “Constitutional AI”: the model checks its answers against a set of principles that the company publishes openly.

There are several models of different sizes — for example Haiku, fast and inexpensive; Sonnet, an all-rounder; and Opus, for hard tasks. New versions come out several times a year. The models can reason before answering, read up to a million tokens at once and use tools: search the web, run code, work with files and connected services.

Keep the limits in mind. Like any neural network, Claude can be confidently wrong, and it knows about events after its training date only if it is given search. By itself it does not remember conversations: memory in the app is notes added to the context. This is an unofficial walkthrough: I am not affiliated with Anthropic and write from its public materials.

Questions and answers

What is Claude?
An AI assistant from the American company Anthropic. You chat with it like a chatbot: ask it to write or edit text, go through a document or spreadsheet, explain a topic, write code. Inside is a large language model that writes its answer one token at a time.
Who made Claude?
Anthropic, a company based in San Francisco. It was founded in 2021 by former OpenAI employees, among them Dario and Daniela Amodei. The company showed the first version of Claude in March 2023 and has released new models regularly since.
How does Claude work?
The app combines system instructions, the conversation so far, files and your question into one text. The model reads this context and writes the answer token by token. On a hard task it reasons first, and for fresh data it turns to tools — web search, for example.
How is Claude different from ChatGPT?
They are products of different companies: Claude is made by Anthropic, ChatGPT by OpenAI. They work in a similar way — both run on large language models. They differ in training data and methods, answer style, tools and prices. Which is handier depends on the task — it is easiest to compare on your own example.
Does Claude remember my conversations?
The model itself does not: its weights do not change after training. Within one chat Claude sees the whole conversation. Between chats there is a memory feature: the app saves notes on what matters and adds them to new conversations. Memory can be turned off in settings.
Does Claude learn from my chats?
On a personal account you decide: the privacy settings have a Model Improvement option. If it is on, chats may be used to train future models — except incognito chats. Conversations flagged for safety review may be used by Anthropic regardless of the setting.
Can Claude be used in Russia?
Not officially: Russia is not on the list of countries and regions where Anthropic provides access to Claude and the API. The current list is on Anthropic’s website, on the Supported countries page.
Is Claude free?
There is a free plan with limits and paid subscriptions with higher limits and extra features. Developers pay for the API by the amount of text — for the tokens the model reads and writes.

I will connect a neural network to your bot or site

I will pick a model for your task, budget and the country your service runs in, and build it into a Telegram bot or website — like my “Что там?” bot, where a neural network builds a topic profile from a couple of lines.