The official YouTube channel for Claude, built by Anthropic. Talks and tutorials by the people who build Claude, Claude Code and the Model Context Protocol — including the "Code w/ Claude" conference sessions many of our lessons are based on.
Visit channel ↗A.A.Murakami create immersive, multisensory installations merging technology with ephemeral natural phenomena like fog, bubbles, and plasma. In 'The Moon Underwater,' Claude serves as a studio collaborator for their creative process.
An Anthropic workshop where participants build diamond-mining agents in 45 minutes and compete on a live leaderboard. Learn agent configuration, eval-driven improvement, and what separates winning architectures.
Agent view in Claude Code provides one place to manage all your Claude Code sessions. Available as a research preview on Pro, Max, Team, Enterprise, and Claude API plans.
Sales reps drown in administrative work — digging through scattered documentation to answering customer emails late into the night. Jared Sires, GTM Product Manager, shares how he went from account prep to customer follow-ups with Claude.
Security teams drown in alerts, jumping between tools and query languages for every investigation. Jackie Bow, Technical Lead on Anthropic's Detection Platform Engineering team, shares how her team uses Claude to unify security signals and speed up threat investigation.
There's an opportunity for marketing teams that adopt Claude Code to spend less time on repetitive execution and more time on what matters. Austin Lau, growth marketer, shows how marketing teams reduce execution overhead with Claude.
Product engineers lose hours toggling between tools and tackling subtasks one at a time. Software engineer Chuma Kabaghe shows how she uses Claude Code to onboard onto unfamiliar codebases in minutes, then stay in the flow throughout development.
Getting data as a PM means pinging a data science team and waiting. Lisa Crofoot, PM, shares how Anthropic's PMs use Claude to query product data and build evals in minutes — no more blocked on data requests.
Before a Claude model ships, a small group of partners tests it, breaks it, and shapes what gets released. What pre-release evaluation looks like from both sides of the partnership.
Part 1 of 3. Learn the basics of building a PowerPoint presentation with Claude. See how to use Claude directly in PowerPoint to build and edit slides that stay on-brand.
Claude now works across your open Outlook, Word, Excel, and PowerPoint files in a single continuous session — triaging email, drafting documents, building models, updating slides, and carrying full context across all.
With the Autodesk Fusion connector, designers and engineers can turn natural language into design actions, iterate without starting over, and stay in flow throughout the design process.
With the new Blender connector, you can debug a scene, build new tools, or batch-apply changes across every object, directly from Claude. For 3D artists who know what they want but not the Python API.
The Claude Code desktop app has been redesigned for agentic coding: many things in flight, and you in the orchestrator seat. A new sidebar lets you run and manage sessions in parallel across repos, and a new notifications center keeps you in the loop.
How Claude Code's remote control feature lets the agent keep working while you step away — and how to check in from mobile, manage long-running tasks, and use it safely.
Code Review dispatches a team of agents on every pull request to catch the bugs that skims miss. When a PR is opened, agents search for bugs in parallel, verify each one to filter out false positives, and post only confirmed issues as inline comments.
Claude Design is a new Anthropic Labs product that lets you collaborate with Claude to create polished visual work like prototypes, slides, one-pagers, and more.
Claude now also connects to the apps you use outside of work — from travel and food to entertainment and home. Ask for what you need, and Claude pulls in the right app.
New features that help you tackle harder tasks with more of the tools you use every day. Claude in Excel handles long-running and harder tasks with improved performance. Claude now plans before acting.
In Claude Cowork and Claude Code, you can now let Claude use your computer to handle tasks. It can point, click, and navigate like you would to do everything from opening and editing files to handling complex workflows.
What CLAUDE.md is, what belongs in it, the global-project-folder hierarchy, and how to keep it accurate as your codebase evolves.
Claude can now work across your open Excel and PowerPoint files in a single continuous session — reading live cell values, building out analysis, updating slides, and carrying full context across both apps.
The Claude mobile app now connects to apps that are fully interactive. Pull up live charts, sketch diagrams, and build shareable assets, all rendered visually right in your conversation.
Claude Security, now available in public beta, scans codebases for vulnerabilities and suggests targeted software patches for human review, allowing teams to find and fix issues that traditional methods miss.
A skill teaches Claude how to do something once. Then, Claude applies that knowledge automatically whenever it's relevant. Skills allow users to define custom behaviors that Claude remembers and applies consistently.
Claude now creates diagrams, interactive tools, and visualizations right in the conversation. Follow a student from choosing a major to getting a city park approved — all built visually, all within the conversation.
See how a fictional company uses Claude, Cowork, and Claude-powered products across several teams to respond to market challenges substantially faster than before. Demonstrates Cowork and Plugins integration.
Dispatch is a new way to work with Claude in Cowork — one continuous conversation that picks up where you left off. Assign Claude a task from your phone or desktop, go do something else, and come back to find it done.
10 purpose-built agent templates for financial services spanning research and analysis, risk and compliance, client operations, and finance workflows. Each agent ships with the context, tools, and guardrails needed for financial work.
Legal teams are asked to do more with less. A question lands about a feature launched months ago, and the context you had when you wrote that contract is gone. Claude Cowork helps legal teams quickly come back up to speed and get answers fast.
Recurring reporting eats a marketing ops team's week: pulling numbers from a half-dozen places, building a detailed review for the team and a summary for leadership. Claude Cowork automates this end-to-end.
Walking into a first customer meeting cold means spending the call getting oriented. Account research lives everywhere: CRM, data warehouse, call recordings, Slack, email, the web. Cowork pulls it all together.
12 one-click plugins bundle workflows for common legal skills and practice areas, including commercial legal, corporate legal, IP, litigation, and more. Purpose-built for legal teams in Claude Cowork.
Cowork is now generally available. Built for how teams work — across the tools they already use, with the controls enterprises need.
Describe a task once. Claude runs it automatically — daily, weekly, or whenever you need it. Scheduled tasks in Cowork save prompts and have Claude execute them on a recurring cadence.
The four-phase workflow that consistently produces correct, reviewable results from Claude Code: why each phase matters, how to prompt for it, and common mistakes that skip critical steps.
Claude Fable 5 designs a complete 3D-printable model in a browser-based CAD editor. The editor itself was also created by Fable 5, including the built-in AI copilot that does the modeling.
Claude Fable 5 autonomously plays Factorio, the factory-building game beloved by engineers, strategizing and building an automated factory on its own.
A fluid simulation coded by Claude Fable 5 where the motion is synchronized to the beat of a classical music EDM remix — which Fable 5 produced using code, having never heard music before.
A timelapse of Claude playing Pokémon FireRed from start to finish using only raw game screenshots — no maps, navigation aids, or extra game-state information. Earlier Claude models needed a computer to help them navigate; Fable 5 does it autonomously using only vision.
Claude Fable 5 built a simulation of the solar system, deriving planetary orbital motion from physics first principles and using it to predict solar eclipses.
Claude powers Figma Make, letting designers turn mockups into working software through natural language. Describe what you want to build and get functional, interactive prototypes.
Claude understands your entire workbook — from nested formulas to multiple tab dependencies. Get explanations with cell-level citations, and update assumptions while preserving formulas.
Learn how to use Claude Cowork to edit your files directly and work in your favorite apps. If it's on your computer, Claude can find and edit it. Describe what you need done, and Claude does it.
A deep dive into the mechanics behind Claude Code: the agentic loop, context window management, permission modes, and the tool set that powers each session.
Claude helps HubSpot's customer success teams analyze calls, identify underlying customer concerns, and draft strategic follow-ups. Moving from surface-level responses to deeper customer insights.
Step-by-step guide to every installation method for Claude Code: npm CLI, VS Code extension, JetBrains plugin, web interface, and API key setup.
Claude Managed Agents is a suite of composable APIs for building and deploying agents at scale. Out of the box: native MCP, tool integrations, memory, and all infrastructure needed to run production agents.
When and why to use Opus 4.8 for autonomous, long-horizon development tasks in Claude Code — stronger reasoning, the /goal command, mobile monitoring, and the cost/capability tradeoff with Sonnet.
Lyft built an AI assistant with Claude that resolves routine support issues in seconds — so human agents can focus on complex cases that require real care and empathy.
What the Model Context Protocol is, how to add MCP servers to Claude Code, how to scope them globally or per-project, and which servers deliver the most value for common development workflows.
Eric Liu from Notion shows how to delegate work to Claude directly inside a workspace. Engineers ship code with it; other teams build websites and presentations. Available in private alpha as part of Claude Managed Agents.
Part 3 of 3. Use Claude in PowerPoint to pull data from other sources and analyze your findings. Connect Excel data to PowerPoint slides for data-driven presentations.
Claude Code on desktop can now start dev servers and preview your running app directly in the interface. Claude sees the UI, reads console logs, catches errors, and keeps iterating — no more switching between terminal, browser, and editor.
Anton Osika built Lovable to let anyone turn a conversation into working software. Learn how trust, craft, and radical accessibility are reshaping who gets to be a builder.
Kay Zhu built Genspark believing AI should free people to follow their hearts. How a father's change of heart about his son's future reveals the deeper mission behind the all-in-one AI workspace.
Max Junestrand is building Legora — an AI-native OS for legal teams. Learn how one of the world's oldest professions is being transformed by language models, and how to build on a rising tide.
Michael Truell started coding at 12 and built Cursor to give everyone the feeling of building without barriers. How an AI-native code editor grew from 15 people to serving 60% of the Fortune 500.
Michele Catasta has spent his career removing the barrier to software creation. Now running Replit with 40 million users, he explains why model power is everything and how AI gives people permission to build.
Scott Wu was world champion competitive programmer at 17 and built Devin — the first AI software engineer. Learn what it takes to build truly autonomous coding agents and why the space is too large for any single product.
Part 2 of 3. Use Claude to refine and elevate slide presentations without leaving PowerPoint. Add new slides, edit individual sections and formatting, and maintain visual consistency throughout.
Slack processes billions of messages a day. Claude powers AI features that help users find what they need: search that answers natural-language questions, conversation summaries, and daily recaps.
Claude Sonnet 4.6 is a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design. Features a 1M token context window in beta.
The thinking behind every lap, visualized with Claude. Pattern of Thought, created with the Atlassian Williams F1 Team, debuts at Monaco. Every race is decided by work no camera catches: the focus, judgment, and split-second thinking of an entire team.
A practical introduction to Claude Code — what it is, how it differs from chat-based Claude, the agentic loop concept, and when to reach for it versus lighter alternatives.
Claude Managed Agents is a suite of APIs for building production-ready agents. You define the tools, environments, and success criteria. Claude works until the job is done. Outcomes, multi-agent orchestration, memory, computer use.
Learn what AI researchers mean when they talk about hallucination in AI models, why it may occur, and tactics you can use to spot this in your conversations.
Focusing on political bias as one type of bias in models — why it may occur, what Anthropic does about it, and tactics you can use to spot this in your conversations.
Behind the scenes of our Atlassian Williams F1 Team partnership film, directed by Dan Tobin Smith. How Anthropic approached storytelling about the human side of AI in high-performance racing.
Claude is joining Atlassian Williams F1 Team as their Official Thinking Partner. From race strategy to engineering, Claude will support how the team thinks, plans, and performs.
Anthropic's Legal Counsel and a Freshfields AI Partner discuss how lawyers are using Claude — from chat Q&A to agentic document creation — and why lawyer judgment remains the rudder in the storm.
How to write effective first prompts in Claude Code: exploration-first strategy, approval modes, common starter tasks, and iterative prompting patterns that get consistent results.
How Omni's CTO Chris Merrick designed a multi-agent analytics system powered entirely by Claude — covering coordinator architecture, tool sizing, the semantic layer as CLAUDE.md analogy, evaluation with LLM-as-Judge, and the critical design pivots that drove 86x token growth.
Frontier models are improving faster than most developers realize. Learn how to read the capability curve, what it means for your product strategy, and when to bet on the model vs. engineer around its current limits.
How Man Group deployed AI-researched trading signals running real capital — and the governed skills framework, core data layer, and compliance model that made it possible.
Four levers that separate real leverage from basic use: writing CLAUDE.md that actually works, wiring in tools with MCP, packaging team knowledge as skills, and running auto mode safely.
A hands-on guide to provisioning Claude in Microsoft Azure AI Foundry, connecting it to MCP servers via Claude Code, and deploying enterprise-grade AI agents — from zero to working code.
Watch a complete feedback app go from empty directory to live URL in a single session — using Claude with Google Cloud, subagents, MCP servers, and skills to handle the full software lifecycle across five distinct roles.
Cut cost and boost intelligence with four platform capabilities every developer should know: prompt caching, tool search, context compaction, and the advisor strategy.
Stand up Claude Code on Amazon Bedrock, teach it your team's conventions with CLAUDE.md and Agent Skills, then graduate to full multi-step orchestration with Lambda and Step Functions.
How Claude Design turns natural-language prompts into interactive prototypes, slides, and production-ready assets — and how to go from idea to shipped product in a single conversation.
Inside Emergent's journey to $100M ARR: how autonomous coding agents built on Claude let non-engineers turn plain language into production software — and what it took to make that work reliably.
Three of Europe's fastest-scaling tech companies made three different bets on Claude. Discover how Delivery Hero built an autonomous agent merging 100+ PRs a day, Doctolib governs AI across its entire healthcare engineering org, and monday.com ships AI directly to users who've never written code.
How capabilities that once required heavy external scaffolding — tool use, context management, code execution, and computer use — moved into the model itself, and how they compose into agents that finish work instead of just starting it.
How to build AI agents that accelerate anti-money laundering investigation, reduce alert fatigue in compliance teams, and reach production safely using in-house MCPs, MCP gateways, and rigorous evals.
Three patterns that transfer from coding agents to legal AI — tool use, verification, and human-in-the-loop — and what Legora had to invent from scratch for document editing, linting, and bulk review.
Patent law demands both deep legal research across millions of documents and genuine technical comprehension of novel inventions. Learn how developers can build reliable AI systems for this uniquely demanding intersection — and what it reveals about legal AI more broadly.
Fabian Hedin, CTO of Lovable, walks through the systems behind a platform that serves 600M+ monthly sessions — the fleet-learning layer that catches coding mistakes, the eval loop that gates every model release, and the engineering philosophy that keeps Lovable itself improving.
Most AI value in organizations is trapped in isolated experiments. Learn how Asana built AI Teammates on Claude Managed Agents to bridge the gap — letting agents ideate, orchestrate, and complete real project management work alongside human teams.
Why building a production-ready agent is harder than it looks, and how Claude Managed Agents handles the infrastructure layer so you can focus on what actually makes your agent valuable.
Turn Claude Code from a reactive assistant into a proactive teammate using Routines — scheduled, event-driven agents that read your repo and open PRs before you've touched your laptop.
Hands-on techniques for testing and comparing Claude models against your use case, so you can make a confident call each time a new release ships.
A hands-on walkthrough of Claude Managed Agents: defining agents and environments, creating sessions, streaming events, using outcomes for iterative self-verification, and monitoring live runs in the developer console.
How Base44 went from a solo founder to 80 engineers after its $80M Wix acquisition — using Claude Code to solve onboarding, code review, QA, and experimentation without building heavyweight processes.
How Metaview built an application-review system whose prompts learn from every recruiter decision — a practical blueprint for automated prompt critique and rewriting.
How Spotify reached 99% AI adoption, built Honk — a background coding agent on the Anthropic Agent SDK — and shifted the engineering bottleneck from writing code to orchestrating it.
How to encode your team's expertise into AI agents using Skills, CLAUDE.md, and the Dreaming feedback loop — so every agent runs with your team's collective judgment, not just the model's defaults.
The workflows Claude Code engineers use to stop hand-holding their AI and start orchestrating it — permission architecture, verification-first design, parallel fanout, and headless automation.
A tour of the latest Claude Code features: routines for async automation, agent view for managing parallel sessions, auto mode for safer delegation, hooks for deterministic control, and the redesigned desktop app.
AirOps's hard-won lessons from shipping Claude agents to non-technical enterprise users: intentional scoping, specialized tools over primitive exploration, and sub-agents for context isolation.
Sessions are isolated by default — agents forget everything when they close. This lesson shows how to wire persistent memory onto your agents and use Dreaming to consolidate and improve what they remember over time.
Fiona Fung, Head of Engineering for Claude Code, shares five lessons from rewriting her team's norms when AI changed where the bottlenecks are — from planning and code review to hiring, onboarding, and org shape.
What Fable 5's capabilities unlock, how dynamic workflows reshape engineering at scale, and what it looks like when a company runs on an AI substrate.
How Claude Code's context window works, when to compact vs clear, and practical strategies for keeping sessions lean and productive.
Without evals you're flying blind — reactive to complaints, unable to verify improvements. This lesson shows how to build code and model graders, run QA loops, and turn subjective quality into something you can act on.
How Replit built VibeBench and the Telescope continuous improvement system to turn overnight eval runs into shipped model upgrades — without a human in the loop.
How Cursor gave cloud agents onboarding, dev environments, and the ability to self-report problems — and what the 'agent experience' means for teams shipping parallel agents at scale.
Anthropic's Applied AI team shares three practices that get more out of longer-running agents: letting Claude interview you instead of writing specs yourself, using HTML over Markdown for richer specs, and embedding verification directly into your artifacts so agents can validate their own work.
Use Claude Code's lifecycle hooks to run formatters, block dangerous operations, and enforce team conventions — every time, without relying on Claude to remember.
The biggest Claude Code platform updates from London 2026: routines that trigger on schedules and webhooks, CI that fixes its own failures, the advisor pattern for frontier-quality at lower cost, and self-hosted agent sandboxes.
Design production memory systems for multi-agent architectures using filesystem-based memory stores, optimistic concurrency, and the dreaming feedback loop.
Claude Managed Agents is the fastest path from prototype to production-ready agent. This lesson walks through the three core primitives — Agent (brain), Environment (hands), Session (the binding) — and shows how to wire them into a working incident-response agent.
Two battle-tested playbooks for prompting work: maintaining and migrating existing prompts, and building agentic loops from scratch using evals to drive every decision.
Master effort levels and adaptive thinking to get the best intelligence-speed-cost trade-off from Claude on any task.
The decision framework for knowing when agent logic belongs in a tool, a skill, or a subagent — illustrated through a live decomposition of a 400-line inventory agent.
How Elicit built AshPL — a Turing-incomplete, purely functional DSL — to make their AI research assistant legible, auditable, and faithfully executable.
Boris Cherny and Cat Wu reflect on Claude Code's first year — what changed about verification, why auto mode beat plan mode, how routines became the killer feature, and where engineering orgs are heading.
Cal Rueb's field-tested playbook for getting consistently great results from Claude Code: context curation, permission strategy, planning, parallel sessions and knowing when to course-correct.
Beyond the hello-world server: why MCP exists, its client–server architecture, the three primitives and who controls them, transports, and where the protocol is heading.
How Claude Code works under the hood, the workflows Anthropic's own engineers use, and how to extend it with memory files, slash commands and headless automation.
Anthropic's Applied AI team shows how to evolve a one-line prompt into a reliable, production-quality prompt — structure, XML tags, examples, giving the model an out, and prefills.
Agents are models using tools in a loop. This lesson covers when to build one, how to prompt it — heuristics, budgets, guardrails — and how to evaluate something that takes hundreds of steps.
Erik Schluntz merged a 22,000-line largely Claude-written change into a production RL codebase. This lesson extracts the discipline that makes that safe: you stop being the code writer and become the system designer and verifier.