How debugsaad/remote-multi-tenant-mcp-server works
This repository snapshot contains product documentation and a PostgreSQL initialization script, but no application source. The graph therefore represents the documented product architecture, not verified implementation modules; component paths are null. The README documents OAuth sign-in, authenticated tool execution, tenant data isolation, auditing, rate limiting, and resumable streams. Connections are limited to those explicit flows.
Components 15
Client and Access
- AI Client
- MCP Gateway
- OAuth Server
Tool Execution
- Rate Limiter
- Session Management
- Tool Dispatch
- Domain Service
- Stream Resumption
State and Persistence
- Redis
- PostgreSQL
- Private Notes and Tasks
Audit and Administration
- Audit Trail
- Activity Review
- Admin Controls
Other components
- User
Connections 17
- AI Client → MCP Gateway: sends tool calls
README.md - MCP Gateway → AI Client: challenges
README.md - AI Client → MCP Gateway: discovers metadata
README.md - User → OAuth Server: signs in
README.md - AI Client → OAuth Server: exchanges credentials
README.md - MCP Gateway → Rate Limiter: checks quota
README.md - MCP Gateway → Session Management: touches session
README.md - Rate Limiter → Redis: consumes tokens
README.md - Session Management → Redis: stores session
README.md - MCP Gateway → Tool Dispatch: dispatches tools
README.md - Tool Dispatch → Domain Service: invokes tools
README.md - Domain Service → PostgreSQL: reads and writes
README.md - PostgreSQL → Private Notes and Tasks: isolates rows
README.md - MCP Gateway → Audit Trail: records invocations
README.md - Audit Trail → PostgreSQL: persists in
README.md - AI Client → Stream Resumption: resumes streams
README.md - Stream Resumption → Redis: replays buffered events
README.md