Overview

Datalore covers data exploration, analysis, visualization, and sharing. Its appeal is the combination: teams can query connected databases, continue analysis in notebooks, and publish reports without splitting those steps across separate products. That breadth is useful when collaboration and handoffs matter; a solo user who only needs a local notebook may not need the team workspace or paid cloud plan.

Key features

Jupyter-compatible notebooks support Python, Kotlin, Scala, and R, with coding assistance. Native SQL cells add syntax highlighting and completion, and can pass query results into pandas DataFrames. This makes Datalore a natural fit for analysis that moves between database queries and Python. Supported connections include Amazon Redshift, Azure SQL Database, MariaDB, MySQL, Oracle, PostgreSQL, and Snowflake.

Notebook users can mount AWS S3 and Google Cloud Storage buckets as folders, and clone or pull public and private Git repositories for use as workspace or notebook resources. These integrations bring data and code sources closer to analysis, without replacing the underlying storage or version-control services.

Real-time collaboration supports shared notebooks with view or edit access, while version history helps teams work with evolving analyses. Notebooks can become interactive data stories and be shared as static or interactive reports. Datalore AI assists with writing Python, SQL, and R code, describing findings, and fixing errors; that is most useful when teams want help within their analysis workflow rather than in a separate tool. GPU compute is supported, and notebooks can be exchanged in formats including .ipynb, .datalore, .zip, .py, .kts, .scala, .r, and .html.

Pricing

Cloud Free: 0.00 USD per free. It includes unlimited notebooks, two notebooks running in parallel, one workspace, one interactive shareable report, 120 CPU S machine hours, and 10 GB of cloud storage. This is a useful way to start or run modest workloads, but the parallel notebook, compute, storage, workspace, and report caps constrain teams with heavier use or multiple shared projects.

Cloud: 35.00 USD per month (billed monthly). Each team member gets 750 CPU L machine hours, 6 Datalore credits, and 20 GB of cloud storage. The per-member allocation gives active teams more room to work than Free, but the monthly cost scales with team size.

On-Premises: custom pricing. It offers a self-hosted or dedicated JetBrains-managed instance, unlimited free viewers, advanced team management, SSO options, and audit logs. It is the relevant tier for organizations that need deployment control or stronger identity and administration features; the self-hosted option requires a 64-bit Linux host and does not support ARM, macOS, or Windows.

Datalore offers a 14-day trial. JetBrains states that deployments are SOC 2 Type II certified and GDPR compliant. On-Premises also includes enterprise-grade support and solution architect consultations.

Platforms

Datalore is available on the web and through API, Linux, and self-hosted deployments. Its hybrid execution model and supported GPU compute give teams deployment and workload flexibility, while self-hosting is bounded by the Linux host requirements.

Who it's for

Datalore is best for data teams that regularly combine SQL access, notebook-based analysis, and collaborative reporting. The Free plan can suit a small or exploratory workflow with modest resource needs; Cloud is a better fit when per-member compute and storage allowances matter. Organizations requiring private or air-gapped environments, custom identity providers, or SSO should consider On-Premises. It is less compelling for people who want only a lightweight local notebook or whose deployment environment is not 64-bit Linux.

Pros and cons

Pros: SQL cells pass results into pandas DataFrames, reducing friction between database work and Python analysis.
Pros: Real-time notebook collaboration and static or interactive report sharing support both joint analysis and broader distribution.
Pros: Cloud storage mounts, Git integration, and several database connectors keep common analysis inputs close to notebook work.
Cons: Free use is bounded by 120 CPU S machine hours, 10 GB storage, two parallel notebooks, one workspace, and one interactive report.
Cons: The 35.00 USD per month Cloud plan is billed monthly per team member's stated allowances, so costs rise with team size.
Cons: Self-hosted On-Premises deployments exclude ARM, macOS, and Windows hosts.

Alternatives

Deepnote is another freemium option with a free plan and support across web, desktop, API, and self-hosted platforms; consider it when that broader platform reach is more important than Datalore's stated database and report workflow.

Hex offers a free Community plan at 0.00 USD per free with any data source, all cell types, small compute, and an agent trial; it may suit readers comparing a free notebook-oriented starting point.

Google Colab has a free plan with no guaranteed compute-unit access and limited GPU access, so it may fit readers who can accept uncertain compute availability.

Kaggle Notebooks is free and provides limited availability of P100 GPU and TPU v3-8, CPU/GPU sessions up to 12 hours, TPU sessions up to 9 hours, and 20 GB autosaved disk; choose it when those free notebook resources suit the workload.

Nextjournal offers a free plan with unlimited private notebooks, data and secrets management, and limited concurrent compute resources; it is an alternative for readers prioritizing private notebooks.

Observable has a free plan with few tokens per day, 1 GB per month of file attachments, and embeds with a watermark; it may suit readers whose needs fit those limits.

JupyterLab is free, open-source software that can be installed locally with pip, making it the alternative for readers who prefer a locally installed notebook environment.

marimo is another option in the notebook software category.

Compare more options in Notebook Software, Data Science Platforms, and Coding Playgrounds.

Verdict

Choose Datalore if your team wants SQL, notebook analysis, collaborative editing, and shareable reporting in one workspace, with a cloud or controlled self-hosted route. Its main reason to look elsewhere is the combination of capped free resources and per-member Cloud billing, or a need to self-host on anything other than 64-bit Linux.

	Datalore	Hexagon-ML
Free plan	Yes	No
Free trial	Yes	No
Paid from	—	—
Open source	No	No
Platforms	api, Linux, self-hosted, Web	Web
Free plan	Yes	—
Supported languages	Python, SQL, R, Scala, Kotlin	Python, R, Julia
Model deployment	—	Yes

Datalore vs Hexagon-ML

Datalore

Hexagon-ML