Label Studio
HumanSignalLabel Studio is a data labeling tool for preparing and correcting training data. Its web interface labels text, images, audio, video, HTML and time series, with labeling screens configured from a tag library and exports to common model formats such as JSON and COCO. Connected ML backends can pre-label tasks for review. HumanSignal builds it and sells Starter Cloud and Enterprise editions beside the free Community Edition.
Openness
4 high confidence- license
- Apache-2.0 (OSI)
- source
- public (HumanSignal/label-studio)
- core features withheld
- yes — HumanSignal's edition comparison lists role-based access control, task assignment, reviewer workflows, annotator agreement and performance dashboards, SSO and on-premise deployment as Enterprise features the Community Edition lacks
Label Studio's Community Edition is Apache-2.0 and runs from the public repository. HumanSignal sells an Enterprise edition, deployable on premises, that adds what a labeling team needs beyond one annotator: reviewer workflows, role-based access, task assignment, agreement metrics and SSO. Those features are not in the open build, which makes Label Studio open core.
- https://labelstud.io/guide/label_studio_compare.html recorded 2026-11-12
The comparison says "Label Studio is available to everyone as open source software (Label Studio Community Edition). There are also two paid editions: Starter Cloud and Enterprise." Its table marks Community without role-based access control, task assignment, reviewer assignment, agreement metrics, annotator performance dashboards, activity logs, SSO and "on-premise availability", all present in Enterprise.
- https://raw.githubusercontent.com/HumanSignal/label-studio/develop/LICENSE recorded 2026-11-12
The root LICENSE is the Apache License, Version 2.0 text, "TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION", with no added term.
- https://raw.githubusercontent.com/HumanSignal/label-studio/develop/pyproject.toml recorded 2026-11-12
The manifest declares name "label-studio", description "Label Studio annotation tool" and license "Apache-2.0".
- https://raw.githubusercontent.com/HumanSignal/label-studio/develop/README.md recorded 2026-11-12
The README says "Label Studio is an open source data labeling tool", gives local, Docker and cloud install paths from the repository, and points to a comparison of what "each edition offers" alongside a free trial of the paid "Starter Cloud edition".
Adoption
2 medium confidenceAdoption is the trailing month of PyPI downloads of label-studio. Docker images, which the README offers as an install path, and the paid cloud editions are not counted.
- https://pypistats.org/api/packages/label-studio/recent recorded 2026-11-12
last_month 27,780, last_week 9,283, last_day 1,687 for the label-studio package.
Capability
3 medium confidenceLabel Studio labels almost any data type, with configurable interfaces, model-assisted pre-labeling and broad export formats. Its free edition sits a step below Snorkel Flow, which adds programmatic labeling and weak supervision, and leaves review and quality workflows to the paid edition.
- https://labelstud.io/guide/label_studio_compare.html recorded 2026-11-12
The feature table marks "Custom ML backends", "Predictions from connected models", "Import pre-annotated data" and "Export data" as available in Community, and "Assign reviewers", "Agreement metrics" and "Active learning loops" as unavailable in it.
- https://raw.githubusercontent.com/HumanSignal/label-studio/develop/README.md recorded 2026-11-12
The README lists "Multi-user labeling", "Multiple projects", "Configurable label formats", "Support for multiple data types including images, audio, text, HTML, time-series, and video", import "from files or from cloud storage in Amazon AWS S3, Google Cloud Storage", and export "to various model formats".
Verified 2026-11-12