Lantern

maintenance

by Lantern Systems, Inc. · since 2023 · also known as lantern_hnsw, Lantern Cloud

Postgres vector extension built on USearch HNSW; development has stalled since late 2024.

vectorrelationalextensionopen source (copyleft)legacy

maintenance. No releases since v0.5.0 (Nov 2024) and no commits since Dec 2024; the project appears dormant. pgvector, VectorChord or pgvectorscale are safer choices.

Overview

Lantern is a PostgreSQL extension from the YC-backed Lantern Systems that adds a lantern_hnsw index type based on USearch, with faster external index builds, in-database embedding generation and BM25 text search through its lantern_extras companion. It was pitched as a quicker alternative to pgvector's HNSW. Since v0.5.0 in November 2024 there have been no releases and the repository has had no commits since December 2024, while pgvector closed most of the performance gap. Lantern Cloud's website is still online, but the project looks dormant; prefer pgvector, VectorChord or pgvectorscale for new work.

Strengths & trade-offs

Strengths

  • +USearch-based HNSW with external (out-of-database) index builds
  • +In-database embedding generation via lantern_extras
  • +Postgres transactions and joins around vectors

Trade-offs

  • −No commits or releases since late 2024
  • −Supports only up to PostgreSQL 17
  • −pgvector now matches most of its performance advantages
  • −AGPL licence

Great for

  • +Maintaining existing Lantern installs while planning migration
  • +Benchmarking USearch-based HNSW against pgvector

Think twice for

  • −New projects
  • −Production systems needing active maintenance

Capabilities

Native or official first-party support · 15 of 25 tracked

✓Transactions
✓Joins
✓Secondary indexes
✓Full-text search
✓Vector search
✓JSON
✓Change data capture
✓Triggers
✓Stored procedures
✓Materialized views
✓Pub/Sub
✓Durable persistence
✓Compression
✓Role-based access
✓Online schema changes
–Geospatial
–TTL / expiry
–Columnar storage
–In-memory
–Encryption at rest
–Time travel
–Branching
–Streaming ingest
–Offline sync
–Runs in WebAssembly

Scale & consistency

Comfortable data size
Medium · ≤ 10 TB
Horizontal scaling
None
Multi-region
No
Transactions
Full ACID
Consistency
Strong
Schema
Strict
Architecture
Single node, Client server
Query languages
SQL

Deployment

Self-hosted Vendor-managed cloud Serverless Embedded Edge Kubernetes operator

Ecosystem

compatible withPostgreSQL
apis & protocolsPostgreSQL wire protocol
client languagesAny PostgreSQL driver

Alternatives to Lantern

Compare all→
pgvectorVector40.0

The standard Postgres extension for vector similarity search with HNSW and IVFFlat indexes.

lantern vs pgvector →
VectorChordVector13.7

Disk-friendly Postgres vector extension using IVF + RaBitQ, the successor to pgvecto.rs.

lantern vs vectorchord →
pgvectorscaleVector13.9

TigerData's Postgres extension adding StreamingDiskANN indexes on top of pgvector.

lantern vs pgvectorscale →
ParadeDBSearch Engine25.1

Elasticsearch-quality BM25 search and fast aggregations inside Postgres via the pg_search extension.

lantern vs paradedb →
USearchVector23.1

Single-header C++ HNSW library with SIMD distance kernels and bindings for a dozen languages.

lantern vs usearch →
Vertex AI Vector SearchVector21.2

Google Cloud's managed ANN service built on ScaNN, now part of Gemini Enterprise Agent Platform.

lantern vs vertex-ai-vector-search →

Lantern FAQ

What is Lantern?

Lantern is a vector database by Lantern Systems, Inc., first released in 2023. Lantern is a PostgreSQL extension from the YC-backed Lantern Systems that adds a lantern_hnsw index type based on USearch, with faster external index builds, in-database embedding generation and BM25 text search through its lantern_extras companion. It was pitched as a quicker alternative to pgvector's HNSW. Since v0.5.0 in November 2024 there have been no releases and the repository has had no commits since December 2024, while pgvector closed most of the performance gap. Lantern Cloud's website is still online, but the project looks dormant; prefer pgvector, VectorChord or pgvectorscale for new work.

Is Lantern open source?

Yes. Lantern is open source. License: GNU AGPL v3.

Does Lantern support ACID transactions?

Yes. Lantern supports full ACID transactions, with strong consistency.

Does Lantern scale horizontally?

Lantern does not scale horizontally; it runs on a single node, so you scale it up rather than out. It is comfortable at medium scale (≤ 10 TB).

How can I run Lantern?

You can self-host it and use the vendor's managed cloud. Managed offerings include Lantern Cloud.

What is Lantern best for?

Lantern is a strong fit for maintaining existing Lantern installs while planning migration and benchmarking USearch-based HNSW against pgvector. Think twice for new projects.

What are the best alternatives to Lantern?

Common alternatives to Lantern include pgvector, VectorChord, pgvectorscale, ParadeDB and USearch. The right choice depends on your workload, scale and hosting; WhichDB's smart match ranks them against your requirements.

Is Lantern still maintained?

Lantern is in maintenance mode. No releases since v0.5.0 (Nov 2024) and no commits since Dec 2024; the project appears dormant. pgvector, VectorChord or pgvectorscale are safer choices.

How popular is Lantern?

Lantern ranks #378 of 844 databases on the WhichDB Score (5.6/100), and #23 among vector databases. The score combines public adoption signals such as package downloads, GitHub activity, job posts and developer discussion.