Skip to content
Get packs

doccano

Open-source tool for humans to label text data for machine learning projects.

GitHub Download v1.8.5
Stars
10.8k
Forks
1.8k
Licence
MIT
Language
Python
Last update
7 months ago
Type
App

What it is

doccano is a web-based tool for annotating text, meaning people add labels to sentences or passages to create training data. It supports text classification, sequence labeling such as named entity recognition, and sequence-to-sequence tasks such as summarization. You create a project, upload data, and label it in the browser, with collaboration and a RESTful API available.

  • Label sentiment in customer reviews to build a classification dataset
  • Mark named entities in text to train an entity recognition model
  • Write summaries for documents to create sequence-to-sequence training pairs
  • Invite several annotators to a shared project and collect their labels

Great for

  • Machine learning teams building labeled text datasets
  • Researchers creating named entity recognition or sentiment datasets
  • Small groups who want to annotate text together in one shared project
  • Developers who want to automate labeling workflows through the RESTful API

Not ideal for

  • People who need to label images, audio, or video, since the README only describes text annotation
  • Users who need a fully managed hosted service, because the README covers self-hosting options and a demo only
  • Teams needing enterprise support guarantees, since the README does not describe any support plans
Step-by-step guide: how to use doccanoInstall, first run, examples, tips and FAQs
doccano alternativesdoccano vs DocsGPT

How to get it

  1. 1

    Install with a package manager

    You'll need: Python 3.8+ · Docker (optional, for the Docker route)

    Install with pip

    pip install doccano

    Initialize the database

    doccano init

    Create an admin user

    doccano createuser --username admin --password pass

    Start the web server

    doccano webserver --port 8000
  2. 2

    Run it

    • Run doccano init to set up the database.
    • Create a superuser with doccano createuser.
    • Start the web server with doccano webserver --port 8000.
    • In a second terminal, run doccano task to handle file uploads and downloads.
    • Open http://127.0.0.1:8000/ in your browser.
  3. 3

    Get the most out of it

    • Set DATABASE_URL and install the postgresql extra if you need PostgreSQL instead of the default SQLite.
    • Use the Docker image with environment variables to preset the admin account.
    • Use the RESTful API to push data in or pull labels out of your own pipelines.
    • Invite collaborators to projects to annotate in parallel.
    • Check the official documentation at doccano.github.io for FAQ topics such as adding users to projects.

Downloads

Source code (.zip)master branch · straight from GitHubRelease v1.8.5Release notes and source

Strengths

  • Covers several common text annotation task types in one tool
  • Can run locally with pip or Docker, with SQLite by default and PostgreSQL optional
  • Released under the permissive MIT license

Trade-offs

  • The README is truncated and gives limited detail on specific annotation features and their setup
  • The demo is hosted on a free Heroku address that may not be reliable for real work
  • Docker Compose setup requires a .env file and some familiarity with Docker

How it compares

ProjectStarsTypeLevelUpdatedBest for
doccano (this)10.8kAppbeginner7 months agoMachine learning teams building labeled text datasets
bootstrap-vue14.4kLibraryintermediate7 months agoDevelopers maintaining an existing Vue 2 app that already uses BootstrapVue
vue-awesome-swiper12.8kLibraryintermediate3.9 years agoTeams maintaining an existing Vue 3 app that already uses vue-awesome-swiper v5 and need a bridge while migrating
DocsGPT18.4kAppintermediate4 days agoTeams that want an internal assistant over their own documents and wikis
fashion-mnist12.9kLearning resourcebeginner4.4 years agoMachine learning students comparing classifiers on a familiar format
hoppscotch80.7kAppbeginner6 days agoDevelopers who want to test REST, GraphQL, WebSocket, or other APIs without a heavy desktop client

Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.