Databricks Notebooks
By Databricks Notebooks · databricks.com
- Free plan
- Windows
- macOS
- Linux
- Web
- Android
- API
What Databricks Notebooks is
Databricks Notebooks is a web-based environment for data and machine-learning work. Its listed capabilities include SQL analytics, streaming ingestion, data lake support, governance catalogs, automated discovery, and lineage across columns, pipelines, and reports. Users can work with tables, branch datasets, and roll back to table snapshots. The product supports separate storage and compute, ACID transactions, autoscaling, GPU accelerators, and private deployment.
Machine-learning capabilities include batch inference, forecasting, natural-language analysis, AI explanations, and root-cause analysis. Supported formats include MLflow, Python, scikit-learn, XGBoost, PyTorch, Hugging Face Transformers, and custom code. Deployment is listed as both cloud and hybrid, with cloud deployment options and both storage models supported. Free Edition costs nothing and includes one serverless workspace with limited compute size and usage. Pay as you go charges for products used with per-second usage granularity, but no price is listed. Committed Use Contracts are also available, with no listed price.
Who it suits
It suits people working with data and machine-learning workflows who need notebook-based work alongside analytics, governance, and lineage capabilities. The free edition is an option for those who can work within limited compute size and usage.
Good to know, and what to watch
- Supports SQL analytics and streaming ingestion.
- Includes lineage across columns, pipelines, and reports.
- Supports autoscaling and GPU accelerators.
- Free Edition includes one serverless workspace.
- Free Edition limits compute size and usage.
- Pay as you go has no listed price.
- Committed Use Contracts have no listed price.
Plans
Every plan and pricePay as you go
Pay as you go
Committed Use Contracts
Committed Use Contracts
Free plan · 14-day trial
Getting started
- Open the Databricks Notebooks website
- Start with the Free Edition at no cost
- Use the included serverless workspace
- Choose pay as you go for usage-based charging
- Consider a Committed Use Contract for usage commitments
Limits to check
The Free Edition provides one serverless workspace and limits compute size and usage.
The Club's review
Databricks Notebooks is a web workspace for analyzing data and building machine-learning workflows. It is best for teams that want notebook work connected to SQL, data-lake storage, and governance. Its breadth is the draw; teams seeking only a simple notebook may find more platform than they need.
Overview
Notebooks brings analytics, machine learning, and data management together around tables. SQL analytics, streaming ingestion, data-lake support, and ACID transactions suit workloads that extend beyond exploratory analysis. Separate storage and compute can help teams manage those resources independently, while table-format support and table-level snapshots, branching, and point-in-time rollback provide ways to manage data changes. This is a broad platform choice rather than a narrow notebook utility.
Key features
Analytics and machine learning
Natural-language analysis, AI explanations, forecasting, and root-cause analysis complement SQL and notebook workflows. Support for MLflow, Python, scikit-learn, XGBoost, PyTorch, Hugging Face Transformers, and custom code gives teams room to work across common model formats. GPU accelerators, batch inference, and autoscaling extend the fit to model workloads with changing compute needs. These capabilities are useful for varied analytics and ML work, though the free edition’s limited compute size and usage can constrain meaningful workloads.
Data governance
A governance catalog, automated discovery, column-level lineage, impact analysis, and lineage across pipelines and reports help teams understand data and its dependencies. That makes the platform more relevant to organizations managing interconnected data assets than to individuals working with isolated files.
Pricing
The Free Edition costs nothing and includes one serverless workspace, with limited compute size and usage. It is a reasonable starting point for constrained work, but its caps make it a poor basis for assuming larger workloads will run without cost.
Pay as you go has no up-front costs and charges for products used, with per-second usage granularity. It suits teams whose usage varies and who prefer charges to track consumption, but spending depends on actual product use. Committed Use Contracts involve usage commitments and may bring discounts and other benefits, with flexible commitments across multiple clouds. They suit teams prepared to commit usage in exchange for potential benefits. Both paid options use custom pricing. The pricing note also includes a 14-day trial.
Platforms
Databricks Notebooks runs on the web. Deployment options include cloud, private deployment, and hybrid deployment, supporting teams with different deployment requirements. Storage can be vendor-hosted or customer-managed, and storage and compute can be separated.
Who it's for
Choose it when a team needs notebooks alongside SQL analytics, streaming, machine-learning workflows, and governance for tables. Its combination of modeling support, lineage, and data-lake capabilities is most useful when those jobs belong in one environment. It is less suitable when the requirement is simply a lightweight notebook or when limited free compute would be a barrier.
Pros and cons
- Broad data and ML coverage: SQL, streaming, model formats, GPU acceleration, and batch inference serve more than one kind of workflow.
- Governance built into the scope: Cataloging, discovery, lineage, and impact analysis help teams track data relationships.
- Flexible usage choices: Per-second usage billing avoids an up-front commitment, while committed contracts offer potential discounts and multi-cloud flexibility.
- Free compute is constrained: One workspace and limited compute size and usage make the free edition a limited option for sustained or larger workloads.
- Paid costs depend on use or commitment: Pay-as-you-go spending varies with product consumption, while committed contracts require usage commitments.
Alternatives
For a narrower catalog or lineage need, compare Data Lineage Tools and Query Engine Software; for broader storage and analysis options, see Data Warehouse Software, Augmented Analytics Software, OLAP Databases, and MLOps Platforms.
Atlas is an open-core alternative with an Apache 2.0 core engine and advanced features under a proprietary model. OpenDataDiscovery Platform is a free, open-source platform. Coalesce Catalog offers a free Developer plan for one user, one project, one environment, and 2,000 capped monthly actions. Grai is fully open source and self-hosted, with users retaining control of data and hosting. Apache Atlas is free open metadata governance software. Collate pairs an Apache 2.0 OpenMetadata core with a Community License for its UI and connectors. Amazon DataZone uses pay-as-you-go billing based on requests, metadata storage, compute, and AI recommendation tokens. Google Cloud Knowledge Catalog has a standard-processing free tier covering the first 100 DCU-hours per month.
Verdict
Databricks Notebooks is a strong choice for teams that need notebook-based analytics and machine learning tied to SQL, streaming, data lakes, and governance. The main reason to choose it is that breadth in one web environment; look elsewhere if you need only a lightweight notebook or cannot accommodate usage-based or committed paid plans after the free edition’s limited compute.
The verdict
Databricks Notebooks spans data analytics, governance, and machine-learning capabilities in a web environment. The free edition has limited compute, while paid options do not list prices here.
Questions people ask
Is Databricks Notebooks free?
Yes. The Free Edition costs nothing and includes one serverless workspace with limited compute size and usage.
Is there a trial?
A 14-day trial is listed.
Which platforms does it support?
The listed platform is web.
Which model formats are supported?
Supported formats include MLflow, Python, scikit-learn, XGBoost, PyTorch, Hugging Face Transformers and custom code.
Does it support private or hybrid deployment?
Private deployment is supported, and deployment is listed as both cloud and hybrid.
What the maker publishes
- AI assistant
- Includes a context-aware Databricks Assistant for conversational data queries. source
- Android app
- Google Play lists Genie One by Databricks as a public-preview Android app with 5K+ downloads and no displayed rating count. source
- App extensibility
- Databricks Apps supports Python or Node.js applications using frameworks such as Streamlit, Dash, Gradio, React, Angular, Svelte and Express. source
- Cell capacity
- A notebook supports up to 10,000 cells, each cell can be 6 MB, and cell output is limited to 20 MB. source
- Code assistance
- The editor provides formatting, autocomplete, Genie Code generation/explanation/debugging and a built-in Python debugger. source
- Collaboration
- Coworkers can edit the same cell simultaneously, see each other's edits and comment on code. source
- Collaboration
- Notebook sharing with permissions, comments and real-time coauthoring. source
- Company founded
- Databricks says it was founded in 2013 by the creators of Apache Spark, Delta Lake, MLflow and Unity Catalog. source
- Company headquarters and size
- LinkedIn lists headquarters in San Francisco and company size of 5,001–10,000 employees. source
- Company legal name
- SEC records identify the company as Databricks, Inc., incorporated in Delaware, with a San Francisco business address. source
- Compliance
- Databricks states offerings for PCI-DSS, HIPAA and FedRAMP and compliance with ISO 27001, ISO 27017, ISO 27018 and SOC 2 Type II. source
- Compute requirement
- A notebook must be attached to a cluster before code cells can run. source
- Execution
- Users can run individual cells, groups or whole notebooks interactively, and schedule notebooks as jobs. source
- Founded
- 2013 source
- Founded
- 2013 source
- Founded
- 2013 source
- Founded
- 2013 source
- Founded
- 2013 source
- Founded
- 2013 source
- Founded
- 2013 source
- Founded
- 2013 source
- Free offering limits
- Free Edition has one serverless workspace and limited compute; no GPUs or enterprise admin. source
- Free plan
- Yes source
- Funding news
- On September 8, 2025 Databricks announced a $1 billion Series K at a valuation above $100 billion. source
Checked against the maker's pages on 18 Nov 2026.