Late-Interaction Vector Database

Delivering better retrieval results for reliable Gen AI

Easily deploy ColBERT in production

Index and retrieve the most relevant documents based on token embeddings.

index = lintdb.IndexIVF("path/to/your/db")
passage = lintdb.RawPassage(
vector_data,
my_document_id
)
index.add(
my_tenant_id,
[passage]
)
Embedded

Work directly with your database in your application.

Streamline development with our embedded database. Remove operational barriers to delivering on production quality.

  • Remove network latency
  • Improve data security and isolation
  • Easier deployment and management
ModelMRR@10
DocT527.6
ANCE33.0
SPLADEv236.8
coCondenser38.2
RocketQAv238.8
ColBERTv239.7

MS Marco dev set scores taken from: ColBERTv2 paper

High Relevancy

Superpowered Gen AI with higher relevancy

ColBERTv2 consistently has the highest relevancy. It's been used to build some of the most advanced Question Answering systems.

Reduce hallucination by incorporating late interaction into your system. LintDB is one of the only vector databases that supports it.

  • Image
  • Audio
  • Text
High Performance

The database for reliable Gen AI

LintDB is the only database that supports late interaction. Increase your search relevancy for improved Gen AI performance.

Embed directly in your application

Simplify Gen AI development with our embedded database. Go to production without the operational concerns.

High Relevancy

Use state of the art techniques to drive better relevance.

Multi-tenancy

Support multiple tenants from one database.

Scalable

LintDB supports merging indexes and read-only replicas. Serve one index from multiple processes.

Fast

LintDB queries your data in milliseconds. Get high performance with top notch retrieval.

Community

LintDB is an open source project powered by individual contributions.

Build reliable Gen AI.

Want more features?

Meet with the Founders