The full engine, on your machine.
Run the complete GPU-accelerated database locally as a Docker container — or spin up free-forever serverless with 10GB of storage and burstable compute up to 4 cores. Vector, geospatial, time-series, and graph, all on your laptop.
Your database, driven by Claude Code.
Install the Kinetica agent skill and Claude Code writes correct SQL, builds applications, and operates the database for you — right inside your local Developer Edition.
Up and running in minutes.
Pull the container, start the engine, and point Claude Code at it. Four steps from an empty machine to a live database.
curl https://files.kinetica.com/install/kinetica.sh -o kinetica && chmod u+x kinetica && ./kinetica start
Requires Docker with at least 8GB of RAM allocated.
The Kinetica shell script manages your installation, starts, and stops the database.
./kinetica --gpu start
./kinetica start
Teach Claude Code to write correct Kinetica SQL, build applications, and interactively operate the database.
npx skills add kineticadb/agent-skills
Workbench is an interactive notebook environment for querying, analyzing and visualizing your data with SQL. You can launch it via your browser using the following link localhost:8000/workbench/
Stop Kinetica
./kinetica stop
Change password
./kinetica alter-password
The default username is "admin." You will be prompted to set a password for this username during the installation process. To change it later, use the command below.
For more commands and configuration options:
./kinetica --help
Stuck on setup? Ask the community.
Having installation issues or a question about the engine? Join the Kinetica community on Slack, or reach out to the team directly.