Skip to content

thamaraiselvam/Routeweave

Repository files navigation

Routeweave

AI-powered API knowledge graph for Node.js backends

Visualise every endpoint, trace business logic step-by-step, and instantly see what breaks when a database column, dependency, or external service changes.

License: MIT Node.js npm


What is Routeweave?

Routeweave scans your Node.js/TypeScript backend with an AI coding agent and builds a rich, interactive knowledge graph. No manual documentation. No API keys required for serving.

Feature Description
Bubble Graph One bubble per endpoint, clustered by route domain (users / tasks / products / …)
API Explorer Searchable list with one-line summaries, flow steps, DB tables, and dependencies
Flow Overlay Step-by-step business logic for each endpoint — with DB, service, and cache nodes
Impact Analysis Select any DB table, dependency, or external service to instantly see every API affected by a change

Demo

Routeweave demo


Quick Start

Point Routeweave at your project and everything happens in the browser:

npx routeweave /path/to/your/repo

Open http://localhost:3789 — a guided setup wizard walks you through three steps:

Step What happens
1. Choose Project Pick an already-scanned repo or point to a new directory
2. Parse & AI Enrichment Static code analysis runs automatically, then an AI coding agent (Claude Code, Cursor, Copilot, etc.) enriches every endpoint with business-logic detail
3. Explore Dashboard is ready — browse APIs, trace flows, run impact analysis

That's it. No extra commands, no config files to edit.


Supported Frameworks

Framework Detection
Express app.get, router.post, method chains
NestJS @Get, @Post, @Controller decorators
Next.js App Router app/api/**/route.ts|js with export function GET|POST…

Cache Files

Routeweave stores scan data in .routeweave/ inside your project:

File Required Description
api_knowledge.json ✅ Required Full API metadata — drives all dashboard views
metadata.json Optional Raw route metadata / audit trail
graph.json Auto-derived Built from api_knowledge.json on first serve
scan_state.json Auto-derived Scan timestamp and API count

Impact Analysis API

GET /api/impact                            # full catalog (tables, dependencies, services)
GET /api/impact?table=users                # all APIs touching `users`
GET /api/impact?table=users&column=email   # APIs that access the `email` column
GET /api/impact?dependency=bcrypt          # all APIs using `bcrypt`
GET /api/impact?service=SendGrid           # all APIs calling `SendGrid`

The dashboard provides a category switcher (Tables / Dependencies / Ext. Services) so you can explore impact across all three dimensions interactively.


Development

Server runs at http://localhost:3789.

TODO

  • Feel free to contribute todo mentioned in docs/TODO.md

Built with ❤️ — scan once, explore forever.

About

Scan once. Explore every endpoint, trace business logic, and see what breaks when your database changes.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors