Spotify

license

Apache-2.0(OSI)

source

public

core features withheld

no

Apache-2.0 license body confirmed. The repository is public and unarchived and builds the whole product, and the README describes no paid tier, enterprise edition or license-gated build beside it, so source is public and the core ungated.

• https://raw.githubusercontent.com/spotify/annoy/main/LICENSE recorded 2026-09-30

  LICENSE file is the Apache-2.0 text: "Apache License" and "Version 2.0, January 2004" and "TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION".

• https://raw.githubusercontent.com/spotify/annoy/main/README.rst recorded 2026-09-30

  README describes "a C++ library with Python bindings to search for points in space that are close to a given query point" and shows no paid tier; searched the body for enterprise, paid tier, license key and commercial: none.

Annoy's adoption reading comes directly from PyPI downloads of the package, since the library itself is the distributed artifact rather than a client for a separate service. That is a direct signal of use, though it counts only PyPI installs and not use through other package managers or vendored copies.

• https://pypistats.org/api/packages/annoy/recent recorded 2026-10-07

  913,200 downloads in the trailing 30 days for annoy (pypistats last_month)

Annoy provides one index family, random-projection-tree indexes stored as mmapped, read-only files with several distance metrics, and an index is immutable once it is built. Chroma covers much more of the same persist-index-retrieve path as a database in its own right, which is why it sits further along the same path.

• https://raw.githubusercontent.com/spotify/annoy/main/README.rst recorded 2026-09-30

  README says Annoy "creates large read-only file-based data structures that are" mmapped so many processes share the data, and that "you can not add more items once the tree has been created".

• https://raw.githubusercontent.com/spotify/annoy/main/README.rst recorded 2026-09-30

  README explains the index is built "by building up a tree. At every intermediate node in the tree, a random hyperplane is chosen" using random projections.

• https://raw.githubusercontent.com/spotify/annoy/main/README.rst recorded 2026-09-30

  README lists distance metrics "Euclidean distance", "Manhattan distance", "Hamming distance" and cosine and dot product, and says Annoy can "use static files as indexes".

• https://raw.githubusercontent.com/spotify/annoy/main/README.rst recorded 2026-09-30

  README says Annoy is "almost as fast as the fastest libraries" and lists only tree-based index building with n_trees, with no other index family; searched the body for HNSW, IVF, quantization, server and cluster: none.