View GitHub

Software

Open source software.

The databases and the tools we build for AI memory. All of it is on GitHub.

View HumemAI on GitHubContact HumemAI
Three stacked blocks with a rose block on top, standing for software built up from parts.

Software

Databases and tools.

Illustration of a document, a triangle graph, vector arrows and a stepped time-series line in one frame around a rose circle at the center, standing for one engine that holds four kinds of data.

ArcadeDB

A database that keeps documents, graphs, vectors, and time series in one engine. Run the official server with one Docker command, and open Studio in your browser.

Learn more
Illustration of three stacked slabs with the middle one in rose, inside a larger outlined square, standing for a database running inside a process.

ArcadeDB in Python

The Python package for ArcadeDB. It runs the engine inside your process, for documents, graphs, vectors and time series, with no service to start. Install it with uv.

Learn more
Illustration for Audit-Ready Memory

Audit-Ready Memory

A public-interest project building local-first, inspectable memory infrastructure for AI systems that need audit trails, deterministic replay, and explicit deletion.

Learn more

Contribute

Everything here is open source. Come build it with us.

The code is public. Open an issue when something is wrong, send a pull request with a fix, or run a benchmark on your own hardware and tell us what you got.

If you work on multi-model databases or agent memory, get in touch.

View HumemAI on GitHubContact HumemAI