Cindy is a media tagging, labelling and querying tool written for use with Machine Learning. It is written in Rust, uses an SQLite database for metadata and is designed to be able to handle large datasets.

Get started

Open-source MIT License. Repository

Security aware

Written in Rust, and designed as a local-only software, Cindy is written with security in mind.

Fast by default ⚡️

Cindy can easily handle up to 100,000 media items in its database.

Simple to use

The simple tag-based categorization and querying makes it very easy to categorize and label even difficult datasets.


Using SQLite as the underlying database for storing metadata, Cindy is fully portable to all systems and datasets are easy to migrate.

Real-time collaboration

Due to the use of WebSockets, Cindy enables real-time collaboration by having all users of the software be fully in sync. Changes made are immediately propagated and shown to other users.


Cindy pulls metadata about files using ffmpeg, which supports a large number of file formats and encodings. It is also possible to extend it in code or using the API.