The least fun part of ML, catalogued

What it does

This repository is essentially a well-organized directory — an awesome-list — of data annotation tools, sorted by modality: images, text, audio, video, time series, 3D imaging, and LiDAR. It links to open-source utilities like labelme and CVAT alongside services such as Labelbox and Dataturks, offering a one-line description for each. The value is in the sorting, not the code.

The interesting bit

The breadth is genuinely cross-disciplinary, stretching from YOLO-specific bounding box editors to Praat for phonetics and webKnossos for neural morphology in 3D. It treats annotation as infrastructure rather than a computer-vision-only chore.

Key highlights

• Covers eight distinct data types, including niche areas like time-series anomaly detection and LiDAR semantic segmentation
• Mix of desktop GUIs, browser-based platforms, and JavaScript libraries for embedding
• Includes specialized scientific tools such as CATMAID for massive collaborative image datasets and KNOSSOS for 3D connectomics
• Entries range from lightweight single-purpose scripts to full collaborative crowdsourcing frameworks

Caveats

• The README offers only a flat list with minimal descriptions; there is no comparison matrix, licensing guidance, or maturity assessment
• Some entries are little more than a link and a tagline, so you will still need to evaluate each tool yourself

Verdict

Worth bookmarking if you are scoping annotation tooling for a new machine learning project and want a modality-sorted starting point. Skip it if you are looking for a framework to import; this is a reference page, not a product.