Label Studio

Label Studio is the most flexible open source data labeling platform for fine-tuning LLMs, training computer vision models, and perfecting audio transcription.

Engineers use Label Studio to build high-quality training datasets across text, image, audio, and video formats. The platform supports complex workflows like RLHF for LLMs and multi-modal classification through a customizable UI and integrated ML backends. With over 15,000 stars on GitHub and support for 40+ data types, it serves as the bridge between raw data and production-ready AI. Teams can automate labeling via pre-predictions from models like YOLO or Segment Anything (SAM) to accelerate development cycles.

https://labelstud.io