Back to Gap Map Infrastructure / Storage & Retrieval

Typesense

open source / Overall score: 3.4

Typesense is an in-memory search engine offering typo-tolerant full-text search with faceting, filtering, geosearch and vector and hybrid search, positioned as a self-hostable alternative to hosted search APIs. It is written in C++ and runs as a single binary or clustered, with a hosted service and paid support sold beside it.

Openness

5 high confidence

5.0

license: GPL-3.0(OSI)
source: public
core features withheld: no

Typesense's GPL-3.0 license is OSI-approved, and the public, unarchived repository builds the whole product with no paid tier or enterprise edition gating it. The README explains the copyleft choice as keeping modifications open rather than limiting who may use the software or at what scale.

https://raw.githubusercontent.com/typesense/typesense/v31/LICENSE.txt recorded 2026-09-30
LICENSE file is the GPL-3.0 text: "GNU GENERAL PUBLIC LICENSE" and "Version 3, 29 June 2007".
https://raw.githubusercontent.com/typesense/typesense/v31/README.md recorded 2026-09-30
README says "The open source version is free to use" and explains the GPL choice under "Why the GPL license?", with Typesense Cloud and paid support offered separately; searched the body for enterprise, license key and paid tier: only paid support and the managed cloud were found.

Adoption

3 low confidence

3.0

The typesense package on PyPI is a Python client for a running server, so its downloads measure client installs rather than deployments of the server itself. With no published deployment count for the server, GitHub stars are the only adoption signal available.

https://api.github.com/repos/typesense/typesense recorded 2026-08-18
Repo metadata - stargazers_count = 26,444 - for typesense/typesense.

Capability

4 medium confidence

4.0

Typesense is a complete in-memory search engine with full-text, vector and hybrid retrieval, faceting, filtering, geosearch and merchandising rules, clustered for high availability with each node holding a full copy of the data. It sits one step below Vespa, which distributes data across a cluster and covers more of the retrieval path.

https://raw.githubusercontent.com/typesense/typesense/v31/README.md recorded 2026-09-30
README lists "Raft-based Clustering:" as a feature with "Setup a distributed cluster that is highly available." beside geo search, vector search, semantic and hybrid search, and curation and merchandizing.
https://typesense.org/docs/guide/high-availability.html recorded 2026-09-30
High availability page says "each node holds the entire replicated copy of your dataset and can handle searches independently".

Verified 2026-09-27