Dolt

Overview

Dolt combines a SQL database with Git-style workflows: users can fork, clone, branch, merge, push and pull data. It can be connected to like a MySQL database or deployed as a versioned MySQL replica. The result is a practical fit for teams that need to inspect and collaborate on data changes, rather than just keep a database current.

Key features

• Version control for tables: Dolt supports branches, commit inspection, diffs, SQL conflict resolution and commit reverts. Dataset branching and point-in-time rollback are supported at table snapshot granularity, which gives teams a way to work on data changes separately and review what changed.
• Web collaboration: DoltHub hosts public data for free and offers editing and pull-request review through a web GUI. Databases are public by default; Pro users and organizations can create private databases with collaborator and organization permissions. That makes the free collaboration option a poor fit for confidential data.
• Hosted and self-managed options: Hosted Dolt includes a web SQL workbench with pull requests, diffs, commit logs, ER diagrams and CSV upload and download. It also provides nightly full database backups retained for 14 days, AWS and GCP private networking, and enterprise support with deployment-console tickets answered by an on-call team member within the hour. DoltLab instead runs in your own network or on a development machine.
• Integrations: Official connector libraries include JDBC, Perl DBI, MySQL Connector/Python and Go sqlx/database/sql. Tested framework and ORM integrations include Django, SQLAlchemy, Prisma, Laravel, Hibernate, GORM, Rails and FastAPI. This breadth makes Dolt more suitable for existing SQL application stacks than a database that requires an entirely new access pattern.

Pricing

The open-source Dolt plan is free at 0.00 USD per free and includes about 100 MB of program capacity. It is the natural starting point for individual users and teams evaluating versioned SQL data, but the stated program size makes it a limited choice for larger workloads.

DoltHub Pro costs 5.00 USD per month, billed monthly. It includes 100 MB of private data storage free, with storage over 5 GB charged at $1/GB/month. Pro is the relevant step up for users who need private databases rather than DoltHub's public-by-default setting; its low entry price does not remove the storage charges for larger private datasets.

The Hosted Dolt trial instance costs 50.00 USD per month, billed monthly, for a t2.medium EC2 instance and 50 GB of storage. Despite the plan name, it has no time limit. It suits users who want a managed deployment with the hosted workbench and operational features, rather than the free open-source program or a DoltHub collaboration account.

DoltHub says it also sells Hosted Dolt, DoltHub Pro and DoltLab Enterprise alongside the free open-source product. There is no free trial.

Platforms

Dolt supports API, Linux, macOS, Windows, web and self-hosted use. That mix supports both locally managed deployments and browser-based collaboration, though the self-hosted route puts operation within your own network or development machine.

Who it's for

Choose Dolt when your work depends on SQL tables but benefits from branching, diffing, rollback and pull-request review of data changes. It is particularly suited to teams already using MySQL-compatible connections or one of its officially integrated connectors and frameworks. Look elsewhere if private data must be shared through DoltHub without moving to Pro, or if a roughly 100 MB free program is too constrained.

Pros and cons

• Pro: Branches, diffs, conflict resolution and commit reverts bring concrete review and recovery workflows to SQL data.
• Pro: Users can choose hosted collaboration, a managed Hosted Dolt deployment or DoltLab in their own environment.
• Pro: The connector and framework coverage spans several established database and application stacks.
• Con: DoltHub databases are public by default, so private collaboration requires Pro or an organization account.
• Con: The free Dolt program is about 100 MB, and private storage costs can rise above 5 GB on Pro.
• Con: The hosted instance starts at 50.00 USD per month, which is a distinct ongoing cost for managed deployment.

Alternatives

For a wider set of data and ML workflow choices, browse Data Version Control Tools.

• Weights & Biases is worth considering when model seats, storage and Weave data ingestion quotas are the more relevant dimensions for your work.
• ClearML is an alternative if you want a self-hosted version described as 100% open source on GitHub.
• DataLad is a free, open-source option for Linux, macOS and Windows users seeking a narrower platform footprint.
• Neon is a fit for those comparing a hosted database free tier with stated project, compute, storage, branch and egress allowances.
• Nile offers a free local option that runs entirely on your machine without a cloud account.
• Roboflow is oriented toward model training and inference, with a free monthly credit allowance.
• Hugging Face Hub may suit users prioritizing model and dataset hosting, with a free user or organization plan that includes 100GB private storage.
• Quilt is another option for package catalogs and data tooling.

Verdict

Dolt is a strong choice for teams that want database changes to be branchable, reviewable and reversible without giving up SQL workflows. Its version-control model and MySQL compatibility are the main reasons to choose it; look elsewhere if public-by-default sharing or the cost of private and hosted options conflicts with your data or budget needs.