Label Studio is the leading open-source platform for annotating training data for ML models.

Installation

pip install label-studio
label-studio

Browser: http://localhost:8080

Features

  • Multi-Format: Text, images, audio, video, PDFs
  • ML-Assisted: Auto-labeling with pretrained models
  • Active Learning: Prioritize hard examples
  • Collaboration: Multi-user projects
  • API: REST API for automation

Labeling Types

  • Text Classification: Label categories
  • NER: Mark named entities
  • Image Boxes: Draw bounding boxes
  • Audio: Transcribe speech
  • Segmentation: Pixel-level labels

Workflow

Create Project → Upload Data → Define Schema
→ Annotate → Export → Train Model

API Example

from label_studio_sdk import Client

ls = Client(url='http://localhost:8080', api_key='key')
project = ls.create_project(title='Project')
project.import_tasks([
    {'data': {'text': 'Sample'}}
])

Sources