Tired of the overhead that comes with setting up a database for small projects or experiments? So was I. I kept running into the same issues: spinning up a Neo4j instance for tiny graph experiments, switching between SQL, NoSQL, and graph libraries, and fighting with frameworks just to test an idea.
That’s why I built Coffy, a pure-Python embedded database engine that ships with three engines in one library. With Coffy, you get a JSON document store with chainable queries, auto-indexing, and local persistence, a graph engine for building and traversing graphs, and a SQLite ORM with models, migrations, and tabular exports. The best part? All engines run in persistent or in-memory mode, with no servers, drivers, or environment juggling required.
Coffy is perfect for rapid prototyping without infrastructure, embedded apps, tools, and scripts, and experiments that need multiple data models side-by-side. It’s not meant for distributed workloads or billion-user backends, but for developers who want to focus on building without worrying about the underlying database setup.
Coffy is open source, lean, and developer-first. If you’re curious, check it out at https://coffydb.org, or install it via PyPI at https://pypi.org/project/coffy/. You can also explore the source code on GitHub at https://github.com/nsarathy/Coffy.