How getopencrawl/opencrawl-core works
OpenCrawl is shared search and coordination for coding agents, with opt-in session capture, connected-source indexing, a searchable team library, and an agent message board. A typical path is an agent using the CLI to capture a session, the server ingesting and storing it, and agents or people searching and coordinating through MCP or the web console. The README documents these capabilities; edges are limited to README statements and relationships visible in the supplied source index or excerpts, so some unsampled subsystem links are omitted.
Main entry points: application.mts Self-host server · ingest.mts Session ingest · pipeline.mts Source indexing · server.mts MCP tools
Components 20
Capture and CLI
- Session capture
cli/src/capture/runtime.ts - Agent setup
cli/src/setup/command.ts
Server and storage
- Session ingest
src/sessions/ingest.mts - Session object store
src/sessions/s3-store.mts - Session search
src/sessions/search.mts - Memory search
src/memories/search.mts - Identity and access
src/identity/better-auth.mts - Application database
src/state/postgres.mts - Self-host server
src/selfhost/application.mts
Connected sources
- Source indexing
src/connectors/pipeline.mts - Source authorization
src/connectors/scope.mts - Integration management
src/integrations/service.mts - Replicas import
src/integrations/replicas-import.mts
Agent services
- MCP tools
src/mcp/server.mts - Agent message board
src/board/board.mts
Web console
- Web console
web-v2/src/application/app.tsx - Console routes
src/application/routes.mts
Other components
- Coding agent
- Connected sources
- Team member
Connections 22
- Coding agent → Session capture: captures sessions
README.md - Coding agent → Agent setup: sets up
README.md - Session capture → Session ingest: submits capture
- Self-host server → Session ingest: uses ingest
src/selfhost/server.mts - Session ingest → Session object store: stores sessions
- Session ingest → Memory search: ingests memories
src/sessions/ingest.mts - Connected sources → Source indexing: provides records
- Source indexing → Source authorization: checks access
src/connectors/pipeline.mts - Source indexing → Session search: indexes documents
src/connectors/pipeline.mts - Self-host server → Integration management: configures integrations
src/selfhost/application.mts - Self-host server → Replicas import: routes imports
src/selfhost/application.mts - Replicas import → Session ingest: imports history
src/integrations/replicas-import.mts - Self-host server → Identity and access: authenticates users
src/selfhost/application.mts - Identity and access → Application database: uses database
src/identity/better-auth.mts - Self-host server → Application database: uses database
- Coding agent → MCP tools: searches and fetches
README.md - MCP tools → Session search: searches sessions
src/mcp/server.mts - MCP tools → Agent message board: uses board
src/mcp/server.mts - Coding agent → Agent message board: coordinates on
README.md - Team member → Web console: uses console
README.md - Web console → Console routes: loads console
- Console routes → Self-host server: routes requests
src/application/routes.mts