SpiderBrain

SpiderBrain adds an understanding layer to your repository.

The app analyzes your codebase and commits a deterministic .spiderbrain/ folder that captures your project's architecture, dependency graph, keystone components, blast radius, and other structural knowledge. It also generates an AGENTS.md file so AI coding assistants can immediately understand how your project is organized.

Unlike documentation, the understanding layer is machine-readable, version-controlled, and travels with your repository.

What this enables

• Commit your repository's understanding alongside your code
• Give AI assistants architectural context before they generate code
• Inspect keystone files and dependency relationships
• Understand what breaks before making changes
• Keep project understanding synchronized as the code evolves

Consume the understanding layer

Anyone can use the open SpiderBrain tooling to consume a committed understanding layer.

npx spiderbrain mcp

Expose the repository to AI assistants through MCP.

npx spiderbrain read

Explore architecture, keystones, dependency maps, and blast radius locally.

POST /v1/answers

Ask deterministic questions such as:

• Why does this file matter?
• What breaks if I change this?
• What are the architectural keystones?

Privacy

SpiderBrain does not publish user's content or data, nor it's own algorithms, only the project's understanding is exported. Public repositories contain only the portable understanding layer needed for consumption. Proprietary scoring also remains private.

Version-controlled understanding

Code explains implementation.

SpiderBrain explains understanding.

Every clone, pull request, AI assistant, and future contributor starts with architectural context instead of rebuilding it from scratch.