Claude feature map

Shared workspaces that give a team a consistent Claude — same system prompt, same knowledge, same tone — across every conversation.

Claude can remember facts about you or your organization across sessions — so it builds on context instead of starting from scratch every time.

Standalone outputs — documents, code, charts — that Claude produces and you can directly use, edit, or download without copying from the chat.

The instructions that shape every response before the user says anything. The highest-leverage thing an operator controls.

Pre-built capabilities you can enable per Project — web search, image generation, code execution, and more. Each Skill changes what Claude can do.

Integrations that give Claude read access to your data sources — Google Drive, Notion, Confluence, and others — without manual copy-paste.

Model Context Protocol — the open standard that lets Claude connect to any tool or data source, not just the ones Anthropic has built.

Claude's ability to call external functions and APIs — so it can take actions, not just produce text. The foundation of everything agentic.

Claude as an active participant in shared workspaces — able to contribute to projects alongside human team members, not just answer questions.

Route tasks to Claude automatically based on triggers or schedules — so work happens without someone manually kicking it off each time.

Claude instances configured to run autonomously within defined boundaries — handling entire workflows end-to-end with minimal oversight.

Claude can control a computer — clicking, typing, navigating interfaces — to complete tasks in applications that don't have APIs.

Claude conducts multi-step research across the web and your documents, synthesizing findings into structured reports — not just a single answer.

Claude works through complex problems step by step before answering — trading speed for depth on tasks that require careful reasoning.

Retrieval-Augmented Generation — Claude answers questions using your documents as its source material, grounding responses in your actual data.

Reuse expensive context across multiple calls — so Claude doesn't re-read a 100-page document every time. Significant speed and cost savings at scale.

Free, Pro, Team, and Enterprise — each tier unlocks different capabilities, context limits, and admin controls. The plan determines what your org can do.

An agentic coding tool that runs in the terminal — writes, edits, runs, and debugs code autonomously within a codebase. Built for developers.

The developer SDK for building custom agents on top of Claude — with structured tool use, multi-turn conversation management, and orchestration primitives.

The unit Claude charges by. Understanding token usage — what consumes them, where waste happens — is the key to keeping costs predictable at org scale.