doccano

doccano is an open-source text annotation tool for machine learning practitioners, developed under the doccano GitHub organization and released under the MIT License. It supports text classification, sequence labeling, and sequence-to-sequence annotation. The latest listed release is v1.8.5, published in January 2026.

What It Is

doccano is a web-based annotation platform for building labeled text datasets. The README describes the workflow as creating a project, uploading data, and annotating, and says a dataset can be built in hours. Its documented use cases include sentiment analysis, named entity recognition, and text summarization.

Annotation Features

The README lists collaborative annotation, multi-language support, mobile support, emoji support, a dark theme, and a RESTful API. A hosted annotation demo is linked from the README. A background task queue handles file upload and download.

Deployment Options

There are three documented ways to run doccano: pip (Python 3.8+), Docker, and Docker Compose. SQLite is the default database, and PostgreSQL is supported through an optional install extra and the DATABASE_URL variable. A nightly Docker tag provides the latest features. One-click deployment is documented for AWS CloudFormation and Heroku.

Setup Path

With pip, users install the package, initialize the database, create a superuser, start the web server, and run the task queue in a second terminal. The Docker Compose setup uses an .env file with admin, RabbitMQ, and PostgreSQL settings. The FAQ covers creating users, adding users to a project, and changing passwords.