Nexus Code Intelligence

Index any GitHub repo. MongoDB stores every symbol and edge. CDC Change Streams sync them to a LadybugDB graph. Each database does what it was built for.

{ }
tree-sitter
Parse real code
💾
MongoDB
Source of truth
CDC Stream
Real-time sync
🕸
LadybugDB
Graph analytics

What This Unlocks

Two databases, one pipeline. Each engine handles what it's best at.

MongoDB Document Storage

Durable writes, replication, auth, backup. find_one({_id}) in microseconds. Full-text search with $text indexes. Aggregation pipelines for analytics.

LadybugDB Graph Traversals

Multi-hop queries in one Cypher call. MATCH *1..5 replaces app-level BFS. Native shortest path. Community detection. Pattern matching.

CDC Always in Sync

Change Streams watch MongoDB writes and sync to the graph automatically. No dual-write complexity. If the graph dies, rebuild from MongoDB in seconds.

Together Polyglot Persistence

Write to MongoDB (safe, scalable). Query the graph for traversals (fast, expressive). Route each query to the engine that handles it best.

1. Indexed Repository

Parsed with tree-sitter, stored in MongoDB, synced to LadybugDB via CDC Change Streams.

GitNexus github.com/abhigyanpatwari/GitNexus Pre-indexed
Loading stats...

2. Benchmark: Who Wins What?

Same data, same questions. Click any test for full details.

Index a repo first, then run the benchmark.