cwork101 / map
Ecosystem map
Anthropic ships Claude as a dozen separate products with a dozen separate docs sites. The relationships between them are rarely written down anywhere. This is that picture.
Where you use Claude
The apps and integrations you actually open. Each is a separate product with its own feature set.
What you configure
The composable layer: context, tools, permissions and delegation. Mostly shared across surfaces.
What you build on
The developer platform and the SDKs, plus the clouds that resell the same models.
What runs underneath
Model families. Every surface above ultimately routes to one of these.
What it costs
The plans and billing models. Which one you are on decides what the layers above are even available.
What governs it
Safety frameworks, usage policy and data handling. These apply on every surface, on every plan.
Hover or focus any entry to trace its documented relationships across layers.
How to read it
Layer 01
Where you use Claude
The apps and integrations you actually open. Each is a separate product with its own feature set.
Layer 02
What you configure
The composable layer: context, tools, permissions and delegation. Mostly shared across surfaces.
Layer 03
What you build on
The developer platform and the SDKs, plus the clouds that resell the same models.
Layer 04
What runs underneath
Model families. Every surface above ultimately routes to one of these.
Layer 05
What it costs
The plans and billing models. Which one you are on decides what the layers above are even available.
Layer 06
What governs it
Safety frameworks, usage policy and data handling. These apply on every surface, on every plan.
Two things the map deliberately does not claim
- The layers are a mental model, not an architecture diagram. Anthropic does not publish one, and several pieces genuinely sit across two layers.
- Connections are drawn from relationships recorded on each page, so a missing line means “not yet documented here”, not “unrelated”.
Documented, not on the map
These entries are covered in the reference but sit below the level of detail the map shows.
- Anthropic Academy
- Claude 101 (Anthropic Academy course)
- AI Fluency: Framework & Foundations
- Auto memory
- Claude Code 101
- AI Capabilities and Limitations
- Introduction to Claude Cowork
- The 4 Properties of AI
- Claude Platform 101
- Fast mode
- Claude Opus 4.8
- Settings and the .claude directory
- Claude Opus 4.7
- Introduction to Model Context Protocol
- Claude Opus 4.6
- Claude Sonnet 4.6
- Sessions
- Claude Sonnet 4.5
- Claude Opus 4.5
- Building with the Claude API
- Claude Mythos Preview
- Context window management
- Retired models
- Claude Code on the web and mobile
- Claude Code in CI — GitHub and GitLab
- Agent view and background agents
- Usage analytics, costs and monitoring