
Infinity
AI-native database server that runs hybrid search over dense and sparse vectors, tensors, full text and structured data.
Infinity is an AI-native database server. It stores tables with vector, tensor, full-text and structured columns, and answers hybrid queries that combine dense and sparse vector search, tensor matching and full-text search with filters. The server runs as a single binary, started as a Docker container or as a system service from its deb or rpm package, and the README also names an embedded mode as a Python module. Applications reach it over the network through a client, which the README installs from pip for Python. In the Docker and binary deployments, client and server run as separate processes.
Worth knowing
The server keeps running as a process, so it needs a host that keeps services alive. A static web space cannot run it. The README’s Docker example uses the nightly image tag, and the documentation banner points to a released version as the latest, so a reader who wants a fixed release has to pick the tag and the documentation version by hand.
Identity
Rendering and deploy
Releases
Adoption
Sources
- 2026-10-28: links, Maintainer
infiniflow.org - 2026-10-28: links, Written in, GitHub stars
github.com/infiniflow/infinity - 2026-10-28: links, Categories, Runtime, Install with
github.com/infiniflow/infinity/blob/main/README.md - 2026-10-28: Runs as, Install with
infiniflow.org/docs/dev/deploy_infinity_server - 2026-10-28: License
github.com/infiniflow/infinity/blob/main/LICENSE - 2026-10-28: Current major version, Last release
github.com/infiniflow/infinity/releases/tag/v0.7.4 - 2026-10-28: First release
github.com/infiniflow/infinity/releases/tag/v0.1.0 - 2026-10-28: Contributors
github.com/infiniflow/infinity/graphs/contributors