Estimate Databricks workload costs in minutes—with transparent assumptions you can review, share, and export.
Lakemeter is an open-source cost estimation and workload sizing tool that runs as a Databricks App with built-in SSO. Configure expected usage, calculate regional costs, and export a detailed Excel estimate instead of maintaining sizing spreadsheets by hand.
Compute and SQL
AI, ML, and data services
- Model Serving
- Vector Search
- Databricks-hosted foundation models
- Proprietary foundation models
- Lakebase
- Databricks Apps
- AI Parse
- Shutterstock ImageAI
See the workload sizing catalog for the inputs and calculation behavior of every workload.
- Transparent calculations — Review usage quantities, billing units, SKUs, rates, VM costs, storage, and discounts behind every total.
- Workload-specific sizing — Use forms tailored to each supported Databricks workload rather than a generic calculator.
- AI-assisted estimates — Describe a workload in natural language, review the suggested configuration, and accept it with one click.
- Excel export — Generate a detailed workbook for customer conversations, procurement reviews, RFPs, and internal planning.
- Cloud and region awareness — Model the cloud, region, tier, and pricing options available in the app.
- Built for teams — Create, duplicate, share, and compare multi-workload estimates in a Databricks workspace.
You need a Databricks workspace and a configured Databricks CLI profile.
git clone https://github.com/databrickslabs/lakemeter-oss.git
cd lakemeter-oss
./scripts/install.sh --profile <your-cli-profile>The one-command installer provisions Lakebase, loads pricing data, deploys the app, and verifies the installation. It typically completes in 5–15 minutes.
For permissions, deployment inventory, non-interactive installation, and troubleshooting, see the installation guide.
Add workload assumptions and calculate their estimated cost:
Export the completed estimate to Excel:
- User guide — Create estimates, inspect pricing, use AI assistance, and export
- Workload sizing guides — Inputs and calculation behavior for every supported workload
- Admin guide — Installation, architecture, permissions, and API reference
- Changelog — Releases and upgrade notes
Bug reports, feature requests, documentation improvements, and pull requests are welcome. Read CONTRIBUTING.md before contributing, or open an issue to start a discussion.
If Lakemeter helps you size or explain a Databricks workload, consider starring the repository so others can discover it.
- Frontend: React, TypeScript, Tailwind CSS, Vite
- Backend: FastAPI, SQLAlchemy, Pydantic
- Database: Lakebase (managed PostgreSQL on Databricks)
- AI: Claude through Databricks Foundation Model APIs
- Hosting: Databricks Apps with SSO and managed compute
Copyright (2026) Databricks, Inc. This software includes software developed at Databricks and is subject to LICENSE.md. Third-party dependency notices are provided in NOTICE.md.