Vector Store Blueprint

Tangle Blueprint for hosted vector storage and similarity search — the missing piece for RAG.

Endpoints

Method	Path	Description
POST	/v1/collections	Create a collection
GET	/v1/collections	List collections
DELETE	/v1/collections/:name	Delete a collection
POST	/v1/collections/:name/upsert	Upsert vectors
POST	/v1/collections/:name/query	Similarity search
DELETE	/v1/collections/:name/vectors	Delete vectors by ID
GET	/v1/collections/:name/stats	Collection stats

Backends

Backend	Use Case	Config
Qdrant	Production RAG	backend = "qdrant", url = "http://localhost:6333"
ChromaDB	Planned	backend = "chromadb"
InMemory	Dev/testing	backend = "inmemory" (default)

Pricing

Dimension	Default	Unit
Storage	0.10 tsUSD	per GB/month
Writes	0.01 tsUSD	per 1K vectors upserted
Queries	0.005 tsUSD	per 1K similarity searches

Build

# Contracts
cd contracts && forge install && forge build

# Operator
cargo build -p vector-store

Run

# Dev mode (in-memory backend)
VSTORE_OP__VECTOR_STORE__BACKEND=inmemory cargo run -p vector-store

# Production (Qdrant)
VSTORE_OP__VECTOR_STORE__BACKEND=qdrant \
VSTORE_OP__VECTOR_STORE__URL=http://localhost:6333 \
cargo run -p vector-store

Quick Start

# Create a collection
curl -X POST http://localhost:8080/v1/collections \
  -H 'Content-Type: application/json' \
  -d '{"name": "docs", "dimensions": 1536, "distance_metric": "cosine"}'

# Upsert vectors
curl -X POST http://localhost:8080/v1/collections/docs/upsert \
  -H 'Content-Type: application/json' \
  -d '{"vectors": [{"id": "1", "vector": [0.1, 0.2, ...], "metadata": {"source": "readme"}}]}'

# Query
curl -X POST http://localhost:8080/v1/collections/docs/query \
  -H 'Content-Type: application/json' \
  -d '{"vector": [0.1, 0.2, ...], "top_k": 5}'