Doccano

Doccano is a free tool for preparing labeled data for machine-learning projects. It suits practitioners and teams who want to run their own annotation workflow, connect scripts through an API, and choose how to deploy and store data. Its breadth is appealing, but SQLite users need to plan carefully around upgrades.

Overview

Doccano brings dataset import, project setup, team annotation, and export into a single workflow. Supported text tasks include classification, sequence labeling, and sequence-to-sequence annotation, with potential applications such as sentiment analysis, named entity recognition, and text summarization. It also lists image and audio/video annotation, model-assisted labeling, and review capabilities.

As an open-source project dating to 2018, Doccano offers deployment choices rather than a managed-product-only approach. That flexibility favors teams comfortable selecting their own installation and database setup; teams seeking a tightly bounded hosted service may prefer a different fit.

For more options, see Data Labeling Software.

Key features

• Configurable workflow: Set up projects, import datasets, add users, annotate records, and export labeled datasets. The sequence makes Doccano useful for teams that need control over the labeling process from intake through output.
• Collaborative annotation and review: Multiple people can work on annotation, and a review workflow is included. This supports shared projects, though the facts do not establish particular approval roles or review stages.
• REST API and model assistance: Scripts can integrate through REST APIs, and the auto-labeling guide demonstrates Amazon Comprehend Sentiment Analysis as well as custom REST API configuration. This makes Doccano a stronger candidate for workflows that pair human labeling with model or script-based assistance.
• Interface options: The project lists multi-language support, mobile support, emoji support, and a dark theme. These are useful accommodations for varied teams and data, without changing the need to choose an appropriate deployment.
• Storage and login integrations: Imported datasets can be stored in Amazon S3 or Google Cloud Storage. OAuth documentation covers GitHub and Active Directory social login and provides Okta setup instructions.

Pricing

Doccano's doccano plan costs free and is an open-source annotation tool installable with pip, Docker, or Docker Compose. There are no paid tiers described. That makes it a straightforward option for users able to handle installation and operations themselves; it does not come with a stated hosted-service tier or seat allowance.

Platforms

Doccano is available through the web and can be self-hosted, with API, Linux, macOS, and Windows support. The installation guide supports Linux, Windows, or macOS machines running Python 3.8 or later. Installation routes include pip, Docker, Docker Compose, source, and cloud deployment. The choice is useful for teams that need deployment flexibility, but requires them to manage the setup they choose.

SQLite 3 is the default database; PostgreSQL is also described, and MySQL is mentioned as an option. The upgrade warning is material: upgrading while using SQLite 3 can result in database loss. Teams relying on that default should account for this risk before upgrading.

Who it's for

Doccano is best suited to machine-learning practitioners and teams creating labeled datasets who value open-source software, collaborative work, API integration, and control over deployment or storage. It is less suitable for users who want to avoid operational choices or cannot accommodate the SQLite upgrade risk in their workflow.

Pros and cons

• Pro: Free, open-source access combined with several installation routes gives technically capable teams flexibility without a paid plan.
• Pro: The project workflow, collaboration, review, and REST API cover both shared manual labeling and script-assisted processes.
• Pro: S3 and Google Cloud Storage support provide options for imported datasets beyond the default database.
• Con: SQLite 3 upgrades can lose the database, so relying on the default carries a concrete maintenance risk.
• Con: Installation and database options place deployment decisions on the user, which may not suit teams seeking a managed service.

Alternatives

• LightlyStudio is another free, open-source option with Linux, macOS, self-hosted, web, and API platforms; consider it when you want to compare another open-source labeling tool.
• Label Studio has a free plan and free trial in a freemium model; consider it if you want to compare a freemium alternative.
• Potato offers a free self-hosted plan with all features included and no paid tiers or usage limits; it may fit users prioritizing those stated terms.
• Roboflow has a free tier with 10 credits monthly, stated as enough for about 30 model trainings or 80,000 inferences; consider it if that credit-based allowance matches your workload.
• Argilla is free, open-source software deployable on Hugging Face Spaces or your own infrastructure; consider it for those deployment options.
• CVAT offers a limited personal and small-team Community plan under the MIT license, plus an online free plan capped at one member, one project, three tasks, and 1 GB; it may suit users whose needs fit those limits.
• Datasaur has a free plan capped at one user, 5,000 labels per year, and 100 MB of storage, with a personal workspace and a Growth trial up to 14 days; consider it if those caps fit a personal project.
• BasicAI offers private-cloud deployment starting at $6,600 a year, with custom seats, storage, model calls, and on-premise deployment options; consider it when those deployment arrangements are a priority.

Verdict

Choose Doccano if you need a free, open-source labeling workflow that combines team annotation, API-based integration, and flexible deployment. Its main advantage is that breadth of control; look elsewhere if you want to avoid self-managed setup or cannot accept the SQLite upgrade risk.