Open Source Collective

Nullcline Labs

Building the infrastructure layer.

We craft high-performance, open-source systems software — the primitives everything else runs on.

Systems software,
built in the open.

Performance First

We write Rust. Every microsecond matters. Our tools are designed for the hot path — zero-copy, lock-free, and allocation-aware.

Open Source

All our work ships under open-source licenses. We believe infrastructure belongs to everyone who builds on it.

Primitives, Not Products

We build the foundational layers — databases, indexes, protocols. The boring stuff that makes exciting things possible.

What we're building

vectors.db

Rust AGPL-3.0

A high-performance, in-memory vector database built from scratch in Rust. Designed for AI/ML workloads that demand speed, flexibility, and correctness.

View on GitHub →
HNSW Indexing
Hierarchical Navigable Small World graphs for approximate nearest neighbor search at scale.
BM25 Search
Full-text keyword search with BM25 scoring, built-in alongside vector similarity.
Hybrid Retrieval
Combine semantic vector search with keyword BM25 — fuse results with reciprocal rank fusion.
Write-Ahead Log
Durable WAL ensures crash recovery. Your data survives restarts without re-indexing.
Encryption
At-rest encryption for sensitive embeddings. Security without sacrificing performance.
Replication & RBAC
Built-in replication for availability, plus role-based access control for multi-tenant setups.
terminal
# Clone and build vectors.db
$ git clone https://github.com/nullcline-labs/vectors.db
$ cd vectors.db && cargo build --release

# Start the server
$ ./target/release/vectors-db --port 8080
  ┌─────────────────────────────────┐
  │  vectors.db v0.1.0              │
  │  Listening on 0.0.0.0:8080      │
  │  HNSW index: ready              │
  │  BM25 index: ready              │
  │  WAL: enabled                   │
  │  Encryption: AES-256-GCM        │
  └─────────────────────────────────┘

# Insert vectors via API
$ curl -X POST localhost:8080/collections/docs/points \
    -H "Content-Type: application/json" \
    -d '{"id": "doc_1", "vector": [0.1, 0.4, ...], "payload": {"title": "Hello World"}}'
{"status":"ok","id":"doc_1","indexed":true}