Claude by Anthropic

The conversational AI assistant at claude.ai, available on web, desktop, iOS, and Android. Claude excels at reasoning, analysis, writing, coding, and research, with web search, file analysis, and shared project workspaces built in.

• Deep reasoning for complex analysis, strategy, and decision support
• 200K-1M token context window for working with large documents
• Shared Projects for persistent context, instructions, and collaboration across your team
• Web search and research capabilities for up-to-date information
• Image understanding for analysing charts, screenshots, and documents

An agentic coding tool available as a command-line interface, in Claude on the web, and in the Claude desktop app. Claude Code lets developers delegate coding tasks, reading, writing, and editing files across your codebase, executing commands, creating commits and PRs, and working autonomously on multi-step development workflows.

• Autonomous multi-step coding: plan, implement, test, and iterate
• Full codebase awareness through file exploration and search
• Git-native: creates commits, branches, and pull requests
• MCP integration for connecting to external tools and services
• Agent teams mode for parallel autonomous sub-agents on large tasks

An agentic tool built into the Claude desktop app that automates complex, multi-step tasks for non-technical users. Cowork can access folders on your computer, plan and execute tasks autonomously, reading, editing, and creating files while keeping you updated on progress.

• Autonomous task execution without step-by-step prompting
• Direct file system access for real document work
• Visual interface designed for non-developers
• Progress updates so you stay informed without micromanaging
• Plugin ecosystem for extending capabilities

Organised folders of instructions, scripts, and resources that Claude loads dynamically to perform specialised tasks. Skills extend Claude's capabilities with domain-specific knowledge and workflows, usable directly in Claude.ai, the Claude desktop app, and programmatically via the Skills API.

• Pre-built Skills for PowerPoint, Excel, Word, and PDF workflows
• Custom Skills for organisation-specific processes and knowledge
• Available across Claude.ai, Claude desktop, and the Skills API
• Dynamic loading means Claude only uses what's needed for each task

An open protocol created by Anthropic for connecting AI models to external tools, data sources, and services. MCP has reached 100M monthly downloads and become the industry standard for AI-tool integration, adopted well beyond just Anthropic's products.

• Universal integration standard: build once, connect to any MCP-compatible AI
• Pre-built servers for Slack, GitHub, Google Drive, Jira, Salesforce, and more
• Standardised authentication and API handling
• Remote MCP connectors for cloud-hosted integrations
• Open-source with broad industry adoption

Anthropic's managed infrastructure for running stateful, long-running AI agents. You define an agent once, and Anthropic runs the agent loop and hosts a secure, per-session workspace where Claude executes code, runs commands, and works with your files. There is no agent runtime or sandbox for your team to build or operate.

• Anthropic runs the agent loop and hosts the execution sandbox, no agent infrastructure for your team to operate
• Persisted, versioned agent configurations, with an isolated, stateful session for every task
• A secure per-session workspace where Claude runs code, executes commands, and works with files
• Built-in Skills and MCP support, file mounts, and a streaming event feed for live progress
• Production-grade, multi-turn agents that hold context across a long-running task