AI Tool Guide

Claude — The Complete Guide

Everything about Claude — Anthropic’s AI assistant known for thoughtful, careful, high-quality responses. How it was built, what makes it different, how to use it well, prompts for every situation, and full technical depth. Three reading levels. Official sources only.

Claude Anthropic ~9,900 words Updated April 2026

What is Claude?

Claude is an AI assistant made by a company called Anthropic. Like ChatGPT, you type something and it replies — but users consistently notice that Claude’s answers feel a little different. More careful. More nuanced. More honest about uncertainty. Better at handling long, complex documents.

You can use Claude for free at claude.ai or on the Claude app on your phone. No download needed for the web version.

One sentence that captures it

If ChatGPT is the confident, fast-talking assistant who always has an answer, Claude is the thoughtful colleague who reads everything carefully, admits when something is uncertain, and writes with noticeably more care.

Who made Claude?

Claude was made by Anthropic, a company founded in 2021. Here is the story of how it came to exist — and it involves a dramatic split from OpenAI.

The people who left OpenAI

In 2020 and 2021, a group of senior researchers at OpenAI became increasingly concerned about the direction of AI development. They felt that as AI became more powerful, safety needed to be the central priority — not an afterthought. The company was growing fast, taking large investments, and there were disagreements about how much to prioritise safety research versus capability development.

In 2021, Dario Amodei — who had been Vice President of Research at OpenAI — left the company. His sister Daniela Amodei, who had been VP of Operations, left with him. They took with them about a dozen other senior researchers and engineers. Together, they founded Anthropic.

Their founding principle was simple but radical: build AI that is safe, and prove that safety and capability are not in conflict — that you can build a very capable AI that is also genuinely trustworthy.

Why this matters to you

The reason Claude behaves differently from other AIs is directly connected to why Anthropic was founded. The people who built Claude genuinely believe that how an AI behaves — whether it is honest, whether it admits uncertainty, whether it avoids harm — is as important as how capable it is. That philosophy is baked into Claude at a fundamental level.

The history of Claude — year by year

2021: Anthropic is founded

Dario and Daniela Amodei co-found Anthropic in April 2021 with initial funding of $124 million from a group of investors. The company sets up in San Francisco. Their first major research effort is developing Constitutional AI — the technique that will eventually define Claude’s character.

March 2023: Claude 1

Anthropic releases the first public version of Claude in March 2023 — the same month OpenAI releases GPT-4. Claude 1 is available through an API and a limited early access programme. It immediately distinguishes itself through longer context windows (the ability to handle more text at once) and a noticeably careful, nuanced writing style. Reviewers note it is more likely to say “I’m not certain about this” than other AI assistants — which some users find refreshing and others find frustrating.

July 2023: Claude 2

Claude 2 arrives in July 2023 with a dramatically expanded context window — up to 100,000 tokens, compared to around 8,000 for most competitors at the time. This means Claude 2 could read and reason about a full novel, a lengthy legal contract, or hundreds of pages of research in a single conversation. For users dealing with long documents, this was transformative.

Claude 2 also improves coding ability significantly and is made available to UK users for the first time.

March 2024: Claude 3 — a family of models

Claude 3 launches in March 2024 as a family of three models — Haiku (fast and lightweight), Sonnet (balanced capability and speed), and Opus (the most capable). This tiered approach lets users choose between speed and power depending on their task.

Claude 3 Opus, at launch, outperforms GPT-4 on several benchmarks including graduate-level reasoning (GPQA), undergraduate knowledge (MMLU), and coding tasks (HumanEval). For many users, this is the moment Claude becomes a serious competitor to ChatGPT rather than an alternative.

Claude 3 also introduces vision capabilities — the ability to understand images as well as text.

June 2024: Claude 3.5 Sonnet

Claude 3.5 Sonnet arrives in June 2024 and immediately becomes the most-used Claude model. It is faster than Opus, considerably cheaper, and actually outperforms Opus on many tasks — particularly coding. In a widely-shared evaluation by Anthropic, Claude 3.5 Sonnet achieves 64% on the SWE-bench Verified coding benchmark, compared to 49% for the previous best. Developers take notice.

This release also introduces Artifacts — a feature in claude.ai that lets Claude create standalone content (code, documents, visualisations) in a side panel that users can interact with separately from the conversation.

2025: Claude 3.5 Haiku, Claude 3.7, and the Computer Use era

Anthropic continues its rapid release cadence. Claude 3.5 Haiku brings the performance of the original Claude 3 Opus to a fast, low-cost model. Claude 3.7 Sonnet introduces extended thinking — the ability to reason through complex problems step by step before producing a response, analogous to OpenAI’s o1. Anthropic also releases Computer Use — the ability for Claude to interact with a computer interface, clicking, typing, and navigating as a human would.

2026: Claude 4

The Claude 4 series, released in 2026, represents Anthropic’s current frontier. Specific architectural details remain confidential per Anthropic’s standard practice, but the models demonstrate strong performance on long-context tasks, complex reasoning, and agentic applications — tasks where Claude operates with some autonomy to complete multi-step goals.

What makes Claude different from ChatGPT?

Both are excellent AI assistants. The differences are real but subtle:

Claude tends to be better at
  • Long documents (books, contracts, reports)
  • Nuanced, careful writing quality
  • Admitting uncertainty honestly
  • Following complex, multi-part instructions
  • Coding (especially debugging)
ChatGPT tends to be better at
  • Plugins and integrations ecosystem
  • Image generation (DALL‑E built in)
  • Voice conversation (GPT-4o)
  • Custom GPTs marketplace
  • Brand recognition / more tutorials online

What can Claude do? Real examples

A student using Claude for the first time

“I had a 40-page research paper to read and summarise for my coursework. I pasted the whole thing into Claude and asked it to: summarise in 5 paragraphs, identify the 3 main arguments, list any weaknesses in the research, and suggest 5 questions for further study. It did all four in one go. What would have taken me three hours took 10 minutes. And the summary was actually good — not just bullet points, but a real understanding of what the paper was saying.”

1. Reading and summarising long documents

Claude’s long context window makes it uniquely suited to working with large documents. Paste in a contract, a research paper, a book chapter, a set of financial statements, or a lengthy report — and ask Claude to summarise, explain, critique, or extract specific information.

Try this now
Here is a document I need to understand: [paste your document]. Please: 1) Summarise it in 3 paragraphs. 2) List the 5 most important points. 3) Identify anything that seems unclear, unusual, or that I should pay particular attention to. 4) Tell me what questions I should ask about this document.

2. Writing with genuine quality

Claude’s writing tends to be more polished and less generic than other AI assistants. It avoids clichés, structures arguments clearly, and adapts tone convincingly. Particularly good for: academic writing, professional reports, persuasive documents, and anything where quality matters more than speed.

Try this now
Write a [type of document] about [topic]. Audience: [describe them]. Purpose: [what you want the reader to think, feel, or do]. Tone: [formal/warm/persuasive/analytical]. The writing should avoid clichés and generic AI-sounding language. Make it sound like it was written by a knowledgeable, thoughtful human.

3. Honest, nuanced analysis

Claude is notably willing to say “I’m not certain,” “there are arguments on both sides,” or “this depends on factors I don’t know.” For analysis and decision-making, this is genuinely valuable — you want honest uncertainty flagged, not confident-sounding guesswork.

Try this now
Analyse [situation, decision, or idea]. I want you to: present the strongest arguments for it, present the strongest arguments against it, identify what I would need to know to make a good decision, and tell me honestly where you are uncertain. Do not just tell me what I want to hear.

4. Coding assistance

Claude is highly regarded by developers for its coding ability — particularly debugging and explaining code. It can write code in most major languages, spot errors, explain what existing code does, and suggest improvements.

Try this now
Here is my code: [paste code]. It is supposed to [describe what it should do]. Instead it [describe the problem]. Please: 1) Identify the bug. 2) Explain why it is causing the problem. 3) Show me the corrected code. 4) Explain the fix so I understand it.

5. Research and learning

Claude is excellent at explaining complex topics at whatever depth you need — and at being honest about the limits of its knowledge. For learning something new, it can act as a patient tutor who gives you increasingly detailed explanations as you ask follow-up questions.

Try this now
I want to understand [topic]. Start with the simplest possible explanation — assume I know nothing. Then, once I confirm I understand the basics, give me a more detailed explanation. After that, tell me what the most important things to learn next would be. Flag anything where experts genuinely disagree.

Getting started with Claude right now

  1. Go to claude.ai on your computer or phone
  2. Click “Sign up” and create a free account
  3. You can start a conversation immediately — type in the message box at the bottom
  4. To upload a document, click the paperclip icon before sending
  5. Claude will respond thoughtfully — if it is uncertain, it will say so
Your first message
Hello Claude. I’m new to AI assistants. I’m a [brief description of who you are]. What are the 3 most useful things you could help me with specifically? Give me a practical example of each one.

Free vs paid

Free tier
  • Claude Sonnet (with usage limits)
  • Long context window
  • File and image uploads
  • Usage limits apply
Claude Pro ($20/month)
  • Access to all models including Opus
  • 5x more usage than free
  • Priority access during busy times
  • Projects and memory features

Source: anthropic.com/claude — April 2026

Getting the most from Claude

Claude responds especially well to clear, detailed instructions. It is designed to follow complex, multi-part prompts precisely — and unlike some AI assistants, it will tell you if your request is unclear rather than guessing and producing something wrong.

What makes Claude respond differently

Claude was trained to be helpful, harmless, and honest — and it takes all three seriously. It will push back on requests that seem harmful. It will admit uncertainty rather than confabulate. It will ask clarifying questions if your request is ambiguous. Working with Claude means working with an AI that treats conversation as a genuine collaboration.

Claude’s unique strengths — and how to use them

Strength 1: Long context — use it fully

Claude’s context window (the amount of text it can hold in one conversation) is among the largest available. Use it to paste in entire documents rather than summarised extracts. The more context you give, the better the output.

Full document analysis
I am going to share a long document with you. Please read it carefully. Then answer these specific questions about it: [list your questions]. When answering, cite specific sections or quotes from the document to support your points. If anything is unclear or missing from the document, flag it.

Strength 2: Multi-step instructions — give it a complete brief

Claude is unusually good at following instructions with multiple components. Give it a full brief rather than asking one question at a time. This produces better, more cohesive output.

Complete brief format
Task: [what you want produced] Audience: [who will read/use this] Purpose: [what it needs to achieve] Tone: [how it should sound] Structure: [how it should be organised] Must include: [specific elements] Must avoid: [specific things to exclude] Length: [approximate word count or format] After completing the task, tell me if anything in my brief was unclear.

Strength 3: Using Projects for ongoing work

Claude Pro includes Projects — a feature that gives Claude persistent memory within a defined project context. You can set background instructions, upload reference documents, and have Claude remember key facts about your work across multiple conversations.

Project setup prompt
This is a new project: [project name]. Here is the background context I want you to keep in mind throughout all our conversations in this project: [background information — your company, your role, your goals, relevant constraints]. In every response, apply this context without me needing to re-explain it. If something I ask conflicts with this context, flag it.

20 high-value prompts for Claude

1. Contract review
Here is a contract/agreement: [paste document]. I need to understand it before signing. Please: summarise what I am agreeing to in plain language, highlight any clauses that seem unusual or unfavourable to me, flag anything I should negotiate or question, and identify what my obligations and rights are.
2. Research paper summary
Here is a research paper: [paste or upload PDF]. Please: state the main research question, summarise the methodology, summarise the key findings, identify the limitations the authors acknowledge, and explain what this means in practice. Write for an intelligent non-specialist reader.
3. Strategic analysis
Here is a strategic situation or decision I am facing: [describe it]. Give me: a structured analysis of the key factors, the main options available and their tradeoffs, what I might be overlooking, the most important risks, and your honest assessment of the strongest path forward. Flag your assumptions.
4. Improve academic writing
Here is a piece of academic writing: [paste text]. Please improve it for: clarity of argument, precision of language, academic tone, logical flow between paragraphs. Do not change my ideas or add new content — only improve how the existing ideas are expressed. Show changes clearly.
5. Code review
Please review this code: [paste code]. Identify: any bugs or errors, potential security vulnerabilities, performance issues, readability problems, and anything that does not follow best practices for [language/framework]. Explain each issue and suggest fixes. Rate the overall quality honestly.
6. Build an argument
I need to argue for the following position: [state position]. Build me the strongest possible case for this position. Include: the 5 strongest arguments, the best evidence or examples for each, and anticipated counterarguments with responses to them. Then honestly tell me the weakest points in my case.
7. Explain any concept at depth
Explain [concept] at three levels: first in one sentence a 10-year-old would understand, then in one paragraph for an educated non-specialist, then in as much technical depth as is genuinely useful for someone who wants to understand it properly. For the technical explanation, use precise language and flag where there is genuine debate or uncertainty among experts.
8. Create a framework
Create a practical framework for [task or decision]. The framework should be: reusable across similar situations, easy to remember, comprehensive but not overwhelming. Present it as a series of questions or steps with a brief explanation of what to consider at each stage.
9. Socratic learning
I want to learn [topic] using the Socratic method. Ask me questions that will help me discover the key concepts for myself rather than just telling me the answers. Start with a question that gets at the fundamental principle, and build on my responses. If I am wrong, guide me gently rather than just correcting me.
10. Honest feedback
Here is something I have created / written / planned: [describe or paste it]. I want honest, critical feedback — not encouragement. Tell me: what is genuinely good, what is weak or needs improvement, what is unclear or confusing, and what you would do differently if this were yours. Be direct.
11. Compare options rigorously
Compare these options: [list options]. Evaluate each one on the following criteria: [list criteria]. Present your analysis in a structured format. Then give me your honest overall recommendation and explain your reasoning. Flag any assumptions you are making.
12. Scenario planning
I am planning to [describe plan or decision]. Help me think through three scenarios: the optimistic case (things go better than expected), the realistic case (things go roughly as planned), and the pessimistic case (things go worse than expected). For each scenario, describe what could happen and what I should do to prepare for it.
13. Write a structured report
Write a professional report on [topic] for [audience]. Structure: executive summary (3 sentences), background, key findings, analysis, recommendations, conclusion. Length: approximately [word count]. Tone: [formal/analytical]. Each recommendation should be specific and actionable, not generic.
14. Prepare for a difficult conversation
I need to have a difficult conversation with [relationship] about [topic]. My goal is [outcome]. I am concerned about [specific worry]. Help me prepare by: suggesting how to open the conversation, anticipating likely responses, preparing what to say for each response, and identifying what I should and should not say.
15. Generate a study guide
Create a comprehensive study guide for [subject/topic] aimed at [level — beginner/intermediate/advanced]. Include: the key concepts to understand, common misconceptions to avoid, the most important questions to be able to answer, suggested learning sequence, and self-test questions at the end of each section.
16. Business plan section
Write the [section name] section of a business plan for [business description]. Target reader: [investor/bank/internal]. Include all relevant sub-sections for this part of a business plan. Be specific and realistic — avoid generic language. Where I need to fill in specific numbers or information, mark it clearly with [brackets].
17. Explain a bill or statement
Here is a bill / statement / notice I received: [paste text]. Please explain it in plain language. What does it mean? What am I being asked to do? Are there any charges or terms that seem unusual? What should I check or question?
18. Extended thinking prompt (Claude 3.7+)
This is a complex problem that requires careful reasoning. Please think through it step by step before giving your final answer. Show your reasoning process, including any assumptions you make, alternative interpretations you consider, and why you ultimately reach the conclusion you do. Problem: [describe problem].
19. Iterative writing improvement
Here is a draft I have written: [paste draft]. Round 1: improve the opening paragraph only — make it more compelling and specific. Once I approve, move to Round 2: improve the structure and flow of the body. Once I approve, Round 3: strengthen the conclusion. Take it one round at a time.
20. Custom persona for a specific task
For this conversation, please act as an experienced [specific expert — e.g. senior product manager, experienced teacher, seasoned investor]. You have [describe background]. You are advising me on [topic]. Give me advice as that person would — specific, experienced, and honest about where they would push back. My situation: [describe it].

Using the Claude API

Claude is available through Anthropic’s API for developers building applications. The API supports all Claude models and can be accessed through the Anthropic SDK for Python and TypeScript, or directly via REST.

Basic API call (Python)
import anthropic

client = anthropic.Anthropic(api_key="your-api-key")

message = client.messages.create(
    model="claude-opus-4-5",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Explain transformer architecture simply."}
    ]
)

print(message.content[0].text)

Source: docs.anthropic.com — Anthropic’s official API documentation

Constitutional AI — the technical foundation of Claude’s behaviour

Claude’s distinctive behaviour — its tendency toward nuance, its honesty about uncertainty, its willingness to push back — is not accidental. It emerges from a specific training methodology called Constitutional AI (CAI), developed by Anthropic and described in a 2022 research paper.

The problem CAI solves

Standard RLHF (Reinforcement Learning from Human Feedback, used to align ChatGPT) requires human raters to evaluate model outputs for harmfulness. This has two problems: it is expensive to scale, and it exposes human raters to harmful content at volume. Anthropic’s CAI addresses both by using the AI itself to perform much of the feedback process.

The two phases of Constitutional AI

Phase 1: Supervised learning with AI feedback (SL-CAI)

A “constitution” — a set of principles — is defined. Anthropic’s constitution draws from sources including the UN Declaration of Human Rights, Apple’s terms of service, and Anthropic’s own research on AI safety. The model is then given a harmful prompt and asked to produce a response. It is then asked to critique that response against the constitution, and revise it. This self-critique-and-revision loop is run multiple times. The revised responses are used as supervised fine-tuning data.

Phase 2: Reinforcement learning from AI feedback (RLAIF)

Rather than using human raters to rank model outputs (as in standard RLHF), a separate “feedback model” is trained to evaluate outputs against the constitution. This feedback model is used to generate preference data — which outputs are better according to the constitutional principles. The main model is then fine-tuned using RL against this AI-generated preference data. Crucially, the constitution’s principles are made explicit to the feedback model, making the value judgements more transparent and auditable than human ratings.

Primary source

Bai, Y., Jones, A., Ndousse, K., Askell, A., Chen, A., DasSarma, N., Drain, D., Fort, S., Ganguli, D., Henighan, T., Joseph, N., Kadavath, S., Kernion, J., Conerly, T., El-Showk, S., Elhage, N., Hatfield-Dodds, Z., Hernandez, D., Hume, T., Johnston, S., Kravec, S., Lovitt, L., Nanda, N., Olsson, C., Amodei, D., Brown, T., Clark, J., McCandlish, S., Olah, C., Mann, B., & Kaplan, J. (2022). “Constitutional AI: Harmlessness from AI Feedback.” Anthropic. arxiv.org/abs/2212.08073

Interpretability research

A distinctive aspect of Anthropic’s research agenda is mechanistic interpretability — understanding what is actually happening inside neural networks, not just what they produce. Their team has published significant work on:

  • Superposition: The finding that neural networks represent more features than they have neurons by overlapping representations — explaining why networks are hard to interpret but also efficient.
  • Circuits: Identifying specific computational circuits inside transformers that perform recognisable operations (e.g. detecting induction patterns, implementing attention heads that copy previous tokens).
  • Features and sparse autoencoders: Using sparse autoencoders to identify interpretable features in Claude — individual directions in activation space corresponding to human-recognisable concepts — as described in the May 2024 paper “Scaling and evaluating sparse autoencoders.”
Primary sources

Elhage, N., et al. (2021). “A Mathematical Framework for Transformer Circuits.” transformer-circuits.pub

Bricken, T., et al. (2023). “Towards Monosemanticity: Decomposing Language Models With Dictionary Learning.” transformer-circuits.pub

Claude model architecture — what Anthropic discloses

Anthropic does not publish the full architectural specifications of Claude models. The Claude model cards — available at anthropic.com — describe intended use, evaluation results, and safety properties without revealing training data composition, parameter counts, or architectural choices in detail. This is consistent with Anthropic’s responsible scaling policy, which ties capability disclosure to evaluated risk levels.

The Responsible Scaling Policy (RSP)

Anthropic’s Responsible Scaling Policy, first published in September 2023, defines a framework of AI Safety Levels (ASL-1 through ASL-4+) with corresponding commitments about what safety research must be completed before deploying models at each capability level. The policy is notable for including specific, measurable commitments — not just principles — including commitments to pause or adjust development if safety evaluations reveal certain capability thresholds have been reached.

Primary source

Anthropic (2023). “Anthropic’s Responsible Scaling Policy.” anthropic.com/responsible-scaling-policy

Extended thinking — Claude 3.7 and beyond

Claude 3.7 Sonnet introduced extended thinking — a mode in which Claude engages in an explicit internal reasoning process before producing its final response. Unlike the hidden chain-of-thought in standard inference, extended thinking makes the reasoning process visible to the user in a collapsible “thinking” section. This is analogous to OpenAI’s o1 approach, and similarly trades inference speed for improved accuracy on complex tasks. Benchmarks from Anthropic’s model card show significant improvements on tasks requiring multi-step reasoning, mathematical problem solving, and complex coding tasks.

Primary source

Anthropic (2025). “Claude 3.7 Sonnet Model Card.” anthropic.com/claude/model-cards

Computer Use

Anthropic’s Computer Use capability, released in public beta in October 2024, enables Claude to interact with computer interfaces — moving a cursor, clicking, typing, and navigating applications. Technically, the model receives screenshots of the current state of the screen and produces actions (described as tool calls) that manipulate the interface. This represents a significant step toward agentic AI — systems that can take sequences of actions in the real world to complete goals, rather than just generating text.

Primary source

Anthropic (2024). “Introducing computer use, a new Claude.ai feature.” Anthropic News. anthropic.com/news/computer-use

API technical reference

The Anthropic API uses the Messages endpoint as its primary interface. Key technical parameters:

Messages API — with system prompt and tools
import anthropic

client = anthropic.Anthropic()

response = client.messages.create(
    model="claude-opus-4-5",
    max_tokens=2048,
    system="You are an expert data analyst. Be precise and cite uncertainty.",
    messages=[
        {
            "role": "user",
            "content": "Analyse the following data: [data here]"
        }
    ],
    temperature=0.3  # Lower = more deterministic
)

# Access the response
print(response.content[0].text)
print(f"Input tokens: {response.usage.input_tokens}")
print(f"Output tokens: {response.usage.output_tokens}")

Model identifiers (April 2026): claude-opus-4-5, claude-sonnet-4-6, claude-haiku-4-5-20251001. Full model reference: docs.anthropic.com/en/docs/about-claude/models