Anthropic writes for the people building Claude. We write for the people using it at work.
The translation layer between Anthropic's docs and getting real work done. No account, no signup, no paywall — open a page and go.
Choose the path that matches how you work. Two minutes to orient.
Turn scattered updates into clear leadership-ready briefs.
Your path: Claude.ai → Projects → Artifacts → Research Mode → Connectors
Ship faster with Claude handling boilerplate, review, and docs.
Your path: Claude Code → Claude.ai for planning → Skills
Turn scattered updates into clear decisions and next moves.
Your path: Projects → Research Mode → Cowork → Team plan
Marketer or content creator? The Creator Track → — rough idea to polished content in one session.
Four prompts that cover 80% of daily work. Copy any one. Paste it into Claude with your content.
Paste any doc, report, or email thread. Get structured summary + action items.
Summarize this into: (1) what matters most, (2) key decisions or action items, (3) anything I should flag. Be specific, not generic.
Paste any draft. Get a cleaner, more direct version that actually lands.
Rewrite this to be clear, direct, and professional. Cut filler. Short sentences. Keep the meaning. Show me the rewrite only, no commentary.
Paste any situation, problem, or decision. Get structured analysis and a recommendation.
Analyze this. Tell me: (1) what's actually going on, (2) risks or gaps I might be missing, (3) what you'd recommend and why. Be direct.
Paste rough notes, bullet points, or a brain dump. Get a polished, structured deliverable.
Turn these notes into a [email / summary / brief / report]. Structured, professional, ready to send. Fill gaps with reasonable assumptions and flag them.
These work in any Claude chat. For consistent results every time, set up your system →
Why Claude matters, in seven cards.
Stop prompting Claude. Start assigning it work.
Constitutional AI — built-in rules, not afterthought guardrails. Safety is architecture, not a patch.
Most natural writer among frontier AI models. Output you can actually use without heavy editing.
Anthropic never trains on Enterprise conversations. Hard policy, not a checkbox. Team and Enterprise inputs/outputs aren't used for training by default. On consumer plans (Free, Pro, and Max), training only happens if you opt in or your conversation is flagged for safety review.
Anthropic's public commitment in their Feb 4, 2026 post "Claude is a space to think." No ads, ever. The argument: ad incentives conflict with honest, helpful answers.
Fable 5 launched June 9 — Anthropic's newest and most capable public model, built for complex, long-running work. Opus 4.8 is now the heavy-duty workhorse for your hardest everyday tasks; Sonnet 4.6 is still the daily-driver default; Haiku 4.5 is still the fast one. Opus 4.8 and Sonnet 4.6 handle up to 1 million tokens of context — hundreds of thousands of words; Haiku 4.5 handles 200,000.
Add-ins for Excel, PowerPoint, and Word are live. Outlook is in beta. Context carries between apps — what starts in chat can end as a deck without copy-paste.
Claude Code is an AI coding agent that lives in your terminal (one of the most starred AI coding tools on GitHub). Skills are reusable playbooks that teach Claude your team's exact workflows. Install community skills in one command.
Four models. Different strengths. Here's the honest breakdown.
The speed demon. Quick drafts, Q&A, email replies, data formatting. Near-frontier intelligence at the fastest speed. 200,000 token context (roughly 150,000 words).
Your daily-driver default on Free and Pro plans. Near-Opus performance on coding, computer use, and office tasks. 1M token context window — now standard, no beta header needed. Supports Extended Thinking.
Anthropic's heavy-duty workhorse for your hardest work — ahead of Opus 4.7 on the toughest coding and agent jobs. The big shift is honesty: Opus 4.8 is far less likely to let flaws in its own code slip by unremarked, and it flags what it's unsure about instead of bluffing. It still self-verifies its output before reporting back. 1M token context, 128k max output. Same $5/$25 per million tokens as Opus 4.7.
Anthropic's newest and most capable model for everyone. Built for complex work that runs for hours, and for telling you what it's unsure about instead of guessing. High-risk technical topics like cybersecurity and biology are automatically handled by Opus 4.8 — by design. 1M token context, 128k max output. $10/$50 per million tokens.
Need the cheat sheet? Get the one-page model picker →
This is where Claude stops acting like a chatbot and starts acting like an analyst.
When you turn on Extended Thinking, Claude reasons through your problem step by step before answering. It shows its work. Better for hard multi-step problems, complex analysis, and anything where getting it right matters more than getting it fast. Available on Sonnet 4.6 and Haiku 4.5. On Opus 4.8 this runs automatically — it decides how deeply to think, and the new effort control lets you dial it up.
Best for: Strategy breakdowns across multiple documents · Legal or policy review at scale · System and architecture planning from multiple angles
| Use when... | Model | Trade-off |
|---|---|---|
| Most daily work: writing, summarizing, Q&A | Sonnet 4.6 | Best speed/quality balance. Default for a reason. |
| Complex reasoning, long agents, coding projects | Opus 4.8 | Slower, costs more. Worth it for hard problems. |
| A genuinely big, multi-hour project across many documents | Fable 5 | Draws down plan limits faster. Save it for big jobs. |
| High volume, fast turnaround, real-time apps | Haiku 4.5 | Fastest and cheapest. Less capable on nuance. |
Plain English rule: If you're not sure, you're using Sonnet 4.6. That's correct.
Sonnet 4.6, Opus 4.8 + Fable 5: 1 million tokens = hundreds of thousands of words = your entire company wiki
Haiku 4.5: 200,000 tokens = roughly 150,000 words = a long book
Prices approximate. Check current pricing at Anthropic →
The quality of the output depends entirely on the quality of the briefing. Every time.
No context. No recipient. No tone. No length. Claude has to guess everything.
Role + context + task + format + tone. Claude has everything it needs.
Same task. Different briefing. This is the whole game.
Context + Task + Format. Master these three and every prompt gets better.
❌ "Write an email about the project."
✅ "You are a project manager updating a client on a delayed launch. The delay is 2 weeks due to vendor APIs missing the SLA..."
❌ "Help me with this data."
✅ "Identify the 3 biggest risks in this plan. Rank by likelihood and impact. Be specific."
❌ "Send me the results."
✅ "3-column table: Risk Name, Impact Level, Mitigation Strategy. Max 1 line per row."
Don't summarize before pasting. Upload entire codebases, email threads, full contracts. Let Claude do the synthesis. That's what it's built for.
Each one uses the Golden Trio. Copy, paste, customize the brackets.
Add a custom instruction like "I'm a designer with little coding experience — explain things simply and make changes incrementally." Claude calibrates instantly. Takes 30 seconds to set up.
— Product Design team
Instead of asking rapid-fire questions, give Claude the full context and say "work through this and commit as you go." Periodic check-ins beat micromanaging every step.
— Security Engineering team
Never paste client PII, passwords, or internal financials. Be careful with links from unknown sources — prompt injection via URLs is a real attack vector. Only connect tools (file systems, APIs, production systems) you understand and trust.
— Anthropic security guidance
In Anthropic's own testing, Claude confidently reported a finding from a test it never actually ran — and only caught it when the user asked "is that actually proven?" One pushback question is the cheapest quality check you have. Use it on anything that matters.
— from a documented limitation in Anthropic's Fable 5 system card
Claude isn't just a chatbox. Here's the full picture of what it can do and where it runs.
Web and mobile app. Chat, Projects, Research Mode, Artifacts, Memory, and Connectors — all in one place. Start here if you're new.
Upload docs, set custom instructions, and every conversation starts with full context. Stop re-explaining yourself every chat.
When Claude generates a doc, table, chart, or code block, it opens as an Artifact you can edit in place. Keep refining without re-prompting from scratch. Stays attached to the conversation.
Reads your entire codebase, writes and tests code, commits directly. one of the most starred AI coding tools on GitHub. Included in every Team plan seat.
claude in your terminal and ask for a code review.The Chrome extension reads console errors, network requests, and live DOM state — so Claude can debug alongside you without switching tabs. Also summarizes pages and drafts replies. Beta for Pro, Team, and Enterprise.
Describe a task — a report, a spreadsheet, a summary — and Claude builds it on your machine. Now with full computer use: Claude opens files, runs apps, points and clicks, and navigates your screen to finish the job. Plus a persistent agent thread on Desktop and mobile (Pro & Max) so you can start a task on your laptop and check in from your phone.
Add-ins for Excel, PowerPoint, and Word that share full conversation context across apps. Skills support and LLM gateway for Bedrock/Vertex users.
Connect Slack, Canva, Figma, Box, and Asana, and Claude can act in them from the chat — draft and preview a Slack message, generate a diagram, turn a conversation into tasks, or pull a file from Box for context. You approve before anything sends. Access depends on your Claude plan and workspace setup.
Build Claude into your own apps. Same models, full control over system prompts, tools, and integrations via the Anthropic API.
Teach Claude your team's workflows once. Install community skills in one command. Like giving a new hire an SOP manual. Team and Enterprise plans now support org-wide Skills management. The new Agent Skills open standard means your Skills work across AI platforms — not just Claude.
Real use cases. No coding required.
A lawyer with no coding background built a predictive text app for a family member with a speaking difficulty — complete with voice output. No engineers involved.
"I'm a developer now." — Anthropic Legal team
Growth Marketing built a workflow that processes hundreds of existing ads, identifies underperformers, and generates new variations automatically. One person operating like a full team.
10x creative output. Zero extra headcount.
Finance team members with zero coding experience now write plain text descriptions of their workflows — "query this dashboard, run these numbers, produce Excel output" — and Claude executes the whole thing.
No IT ticket. No waiting. Just results.
Voice mode: Hold spacebar and talk to Claude instead of typing. Supports 20 languages. Desktop and mobile.
Features and tools that are live right now — verified, translated for humans. Sorted by what you actually have to do about them.
Availability may vary by enterprise plan.
You're already benefiting from these if you use Claude.
Right next to the model picker in claude.ai and Cowork, there's now an effort control. Turn it up when the answer really matters and you want Claude to think more deeply; turn it down for quick questions, which also goes easier on your usage limits. Available on every plan.
Launched June 9 — Anthropic's newest and most capable public model, built for complex, long-running work that can stretch over hours. Opus 4.8 stays the workhorse for your hardest everyday tasks. Fable 5 is the new top of the lineup.
See the guide →Launched April 17. Describe a slide deck, prototype, or one-pager — Claude builds it. Learns your brand from your codebase or uploaded files. Exports to PDF, PPTX, or Canva. Research preview for Pro, Max, Team, and Enterprise.
Claude now remembers across chats for all users. For structured work that needs exact context control, Projects still wins — but casual chats carry forward automatically.
Claude can now generate charts, diagrams, and data visualizations inline in its responses. No exports, no external tools — paste a dataset or describe what you want, and the visual renders in the chat. Great for quick analysis, meeting prep, and one-off dashboards.
See the guide →Cowork hit general availability April 9 on the Claude Desktop app for macOS and Windows. Schedule recurring tasks. New Customize section combines Skills, Plugins, and Connectors. Enterprise gets Analytics API, usage dashboards, OpenTelemetry.
See the guide →Launched April 28. Claude now plugs into the tools creatives already use — Adobe, Blender, Autodesk Fusion, Ableton, and Affinity by Canva, plus four more. Brief Claude in chat, get help inside the app.
Claude reads messy spreadsheets, dense charts, screenshots, and scanned PDFs far more accurately now — useful when the clean source file doesn't exist and all you've got is a picture. Claude also holds long, multi-step work together across a session, so you don't have to re-explain the task halfway through.
Before handing back a finished task, Claude reviews its own work — checking the result against what you actually asked for, and catching its own mistakes first. You get a more finished draft instead of babysitting every step.
Claude for Excel, PowerPoint, and Word is now generally available for paid subscribers. Claude for Outlook is in public beta. Context carries between apps — what starts in a model ends in a deck without re-explaining.
AllTrails, Instacart, Audible, Tripadvisor, Uber, and Spotify joined the connectors directory. These are personal-account connectors. Useful for life admin; check with your IT team before connecting anything to a work account.
Pro, Max, Team, and Enterprise plans got their Claude Code five-hour limits doubled. Peak-hour throttling is gone for Pro and Max. Powered by a new SpaceX compute deal.
Skip this section if you only use Claude.ai for writing and thinking.
Claude breaks a huge coding job into many parallel sub-agents, then checks the work before reporting back. It plans the job first, runs the pieces at once in one session, and verifies the result. Research preview on Enterprise, Team, and Max.
The Messages API now accepts system entries inside the messages array — so you can update Claude's instructions partway through a task instead of starting the conversation over.
Microsoft customers can access Claude Sonnet 4.5, Haiku 4.5, and Opus 4.1 in Microsoft Foundry without a separate vendor contract. Claude also powers the Researcher agent in Microsoft 365 Copilot, and Excel Agent Mode supports Claude in preview.
New sidebar for every active and recent session. Integrated terminal, in-app editor, rebuilt diff viewer, drag-and-drop panes, side-chat shortcut (Cmd+;) for branching off a task without polluting the main thread.
Bundle a prompt, a repo, and connectors into one config that runs on schedule, on API call, or on a GitHub event. Runs on Claude Code's web infrastructure — laptop doesn't have to be open. Research preview for Pro, Max, Team, Enterprise.
Launched April 8 in public beta. Anthropic-managed sandbox, built-in tools, server-sent event streaming. Use the managed-agents-2026-04-01 beta header. Paired with Routines, this is the full stack for hands-off agent work.
Opus 4.7 added an xhigh effort level, and Auto mode rolled out to Max. The /ultrareview command runs a dedicated review session that flags what a careful reviewer would catch.
Deploy Skills across your whole org. Partner Skills directory is live. The Agent Skills open standard means Skills work on other AI platforms, not just Claude.
See the guide →Launched May 11. Run claude agents to get one screen for every Claude Code session — running, blocked, or done. Reply inline to unblock without leaving the dashboard. No more juggling tmux panes. Available on Pro, Max, Team, Enterprise, and Claude API plans.
Launched May 11. The new /goal command lets you define a completion condition and Claude keeps working across turns until it's met. Live overlay shows elapsed time, turn count, and token usage. Works in interactive, -p, and Remote Control modes.
You can't use these directly, but they shape what Claude becomes.
From Anthropic's Feb 4, 2026 post "Claude is a space to think": Claude will never show ads. No sponsored answers, no advertiser influence on what it tells you. When Claude recommends a tool, it's because it's the right one — not because someone paid.
Launched May 5. Ten pre-built finance agent templates (pitchbooks, KYC, month-end close). Moody's MCP integration brings credit and risk data on 600M+ companies inside Claude. Plus eight new data connectors: Verisk, Third Bridge, Fiscal AI, D&B, Experian, GLG, Guidepoint, IBISWorld.
Get the monthly ClaudeHQ update — new features, translated for humans.
No spam. Monthly Claude updates only.
Describe what you want Claude to do, and get a valid SKILL.md template instantly.
→ Want ready-made skills? Browse the free ClaudeHQ Skills Store
Type a goal below — like "Help me plan sprints" or "Write weekly status reports" — and we'll generate the YAML frontmatter block you need.
Quick note: this is a starter, not a finished skill. It works best on common tasks like status reports, briefs, and summaries. For complex or unusual workflows, use Anthropic's skill-creator inside Claude for a guided build.
This is the contents of your SKILL.md file. To use it:
SKILL.md.If your skill doesn't upload, check the filename first: it must be exactly SKILL.md.
The description field controls WHEN Claude loads your skill. Get it wrong and your skill never fires.
Formula: What it does + When to use it
✘ Bad: "Helps with projects."
✔ Good: "Sprint planning for agile teams. Use when user says 'plan sprint', 'create sprint tasks', or needs to organize upcoming work into a two-week cycle."
Write it the way you'd actually ask Claude — not the way a developer would document it.
From Anthropic's official skills guide — three things skills are built for:
1. Document & Asset Creation
Consistent output: docs, presentations, code, designs. No external tools needed.
Example: a skill that always writes proposals in your brand voice.
2. Workflow Automation
Multi-step processes that need to run the same way every time. Often uses MCP servers.
Example: pulls Linear tasks, writes standup summary, posts to Slack.
3. MCP Enhancement
Adds expertise on top of a connected tool.
Example: doesn't just access Sentry — knows your team's bug triage process and applies it automatically.
1. Zip your skill folder (the folder containing SKILL.md)
2. Open Claude → Settings → Capabilities → Skills
3. Click "Upload skill" → select your zip file
4. Test it: type a message that should trigger it and watch for the skill icon
For teams: Admins can deploy skills org-wide from the admin dashboard.
Anthropic built a skill that builds skills. It's called skill-creator — it's free, it's inside Claude, and it walks you through every step in 15-30 minutes.
How to use it:
Option 1 — Claude.ai: Open a new conversation and type:
"Use the skill-creator skill to help me build a skill for [your goal]"
Option 2 — Claude Code: The skill-creator skill is available in your skills directory. Run it from Claude Code for programmatic output.
Check the filename is exactly SKILL.md — case-sensitive (capital S, K, I, L, L, .md). Verify YAML frontmatter has both opening and closing --- delimiters.
Description is too vague. Rewrite it using phrases you'd actually say to Claude. Debug move: ask Claude "When would you use the [skill-name] skill?" — it quotes your description back. Adjust from there.
Add scope to your description. Example: "Do NOT use for general questions — only activate for [specific context]."
Test the MCP independently before blaming the skill. Ask Claude to use the tool directly without the skill active. If that fails, the issue is the MCP connection, not your skill. Check: Settings → Extensions → your service → should show "Connected."
ClaudeHQ is an independent project, not affiliated with Anthropic. Skills content translated from Anthropic's official documentation. All Claude trademarks belong to Anthropic.
The best of Anthropic's own documentation, curated so you don't have to dig.
Two API deadlines hit June 15 and Aug 5. Check what's migrating and when.
Full API and platform documentation
Feature guides and FAQs
Official announcements and research
35 free prompt skills, copy and go
FREE PDF GUIDES
Two API deadlines hit June 15 and Aug 5. Migration pairs, 5-minute checklist, and a plain-English "do I need to care?" decision tree.
The 3-part framework that fixes 80% of bad Claude outputs.
The exact wording errors that make Claude sound generic — and the replacements.
Step-by-step: go from new account to a working Project in under 10 minutes.
All 6 models in the claude.ai picker, with what each is for and when to skip it. Built to screenshot.
Haiku vs Sonnet vs Opus — the honest decision guide with no jargon.
The #1 question new Claude users have. One cheatsheet answers it.
The safety rules your whole team should read before using Claude for work.