Embeddings

Vector search, built in

Curiosity includes a vector database. Embeddings are generated, stored, and queried in the same system as your graph: no separate infrastructure to build or maintain.

Book a Demo

Check Documentation

How it works

From data to semantic search

Mark a field as vector-indexed and Curiosity handles the rest: embedding, indexing, and retrieval all run inside the same platform.

Embed

Text fields are encoded into vectors automatically at ingestion, using your choice of embedding model.

Store

Vectors live in an in-memory index inside Curiosity. No external vector database needed.

Query

Search by meaning, find similar items, or ground LLM responses — all through the same API you already use.

Capabilities

Embedding capabilities

Everything you need for vector search and semantic retrieval, integrated into one system.

Vector Database

Curiosity includes a vector database. No separate infrastructure needed.

Graph Embeddings

Generate embeddings from graph structure and relationships

Custom Embeddings

Integrate externally trained models like domain-specific embeddings

Embedding Pipelines

Run embedding models automatically when new data is ingested

Fast Embeddings

Process large datasets and run real-time inference with low infrastructure overhead

Embedding Setup

Configure chunking, overlap, and indexing settings per data type

Vector Search

Retrieve results by semantic similarity using an integrated HNSW index

Hybrid Search

Combine keyword and embedding search for more accurate results

Similarity Search

Search by semantic similarity with support for mixed retrieval strategies

Graph Boosting

Boost retrieval results based on graph relationships and proximity

Graph Filtering

Filter retrieval results using graph structure and properties

Tech Domain Search

Use graph context to improve accuracy on part numbers and codes

Embeddings

Vector search, built in

How it works

From data to semantic search

Capabilities

Embedding capabilities

Everything you need for vector search and semantic retrieval, integrated into one system.

Developers

Questions for developers

Which model does Curiosity run?

What can I build on Curiosity?

How does the knowledge graph work?

Can I run models on my own system?

Can I connect custom data sources?

How do I build agents?

Can I extend the platform with custom logic?

Can I run Curiosity on my own infrastructure?

How are permissions enforced in what I build?

How do search and embeddings work together?

How do I get started?

Connected knowlege for AI systems

Connected knowlege for AI systems