🔥 Key Features

• 🤖 Built for Agentic AI — Connect AI agents and LLM tools directly through a scriptable CLI
• 🧠 Built-in AI Functions — Leverage native AI/ML capabilities with GPU acceleration, right where your data lives
• 🏎️ In-Memory Performance — Run complex analytics at in-memory speed with Exasol's industry-leading analytics engine
• 🏢 Full Enterprise Features — Access all enterprise-scale capabilities of the Exasol database, completely free for personal use
• ♾️ Unlimited Data — Store and analyze unlimited amounts of data with no artificial limits
• 📈 Scalable Architecture — Scale up to any number of nodes using Exasol's MPP (Massively Parallel Processing) architecture
• ⚙️ Simple Deployment — Spin up Exasol on AWS, Azure, Exoscale, STACKIT, or your local system with just a few commands
• 🖥️ Cross-Platform CLI — Install and manage your cluster using the Exasol Launcher on Linux, macOS, or Windows

✅ Prerequisites

A cloud account on one of the supported platforms with permission to provision compute instances, or a macOS system for local deployment:

• AWS — Set up an AWS account for Exasol Personal
• Azure — Set up an Azure account for Exasol Personal
• Exoscale — Set up an Exoscale account for Exasol Personal
• STACKIT — Set up a STACKIT account for Exasol Personal
• Local — local deployment on macOS, with at least 8 GB RAM

🏎️ Quick Start (macOS / Linux)

• Download the launcher

curl https://www.exasol.com/install/ | sh

• Install on a cloud provider or your local system

exasol install aws        # Amazon Web Services

exasol install azure      # Microsoft Azure

exasol install exoscale   # Exoscale

exasol install stackit    # STACKIT

exasol install local      # local system, macOS

Read on for Windows instructions and full details.

🚀 Deploy Exasol Personal

• Download and install the Exasol Launcher for your platform.

On Linux and macOS, run:

curl https://www.exasol.com/install/ | sh

This installs the exasol binary to ~/.local/bin. On most Linux distributions this directory is already in your PATH. On macOS, or if the installer reports that ~/.local/bin is not in your PATH, follow its instructions.

On Windows: download the Exasol Launcher from the Exasol Download Portal and copy the exasol binary to a directory in your PATH.

• For cloud presets, configure authentication for your provider. See the relevant account setup guide in Prerequisites for the environment variables and credentials required.

• To install Exasol Personal, run the following command with the preset for your cloud provider or local system:

exasol install aws        # Amazon Web Services
   exasol install azure      # Microsoft Azure
   exasol install exoscale   # Exoscale
   exasol install stackit    # STACKIT
   exasol install local      # local system, macOS

The exasol install command does the following: - Generates backend files in the deployment directory - Provisions the necessary deployment resources - Starts the deployment - Downloads and installs Exasol Personal on that infrastructure

The whole process normally takes about 10 to 20 minutes to complete.

When the deployment process has finished, you will see instructions on how to connect to your Exasol database using a client of your choice. You can also find this information at any time by using exasol info in the terminal.

By default, Exasol Personal stores deployment state in ~/.exasol/personal/deployments/default. If you run a command from an existing deployment directory, Exasol Personal uses that directory instead. Pass --deployment-dir <path> to choose a different deployment directory explicitly.

Keep the deployment directory until deployment resources have been destroyed. Deleting the directory does not remove those resources and can make cleanup harder.

An initialized deployment directory is tied to the selected infrastructure and installation presets. Rerun exasol install <preset> with the same presets to retry a failed deployment safely, or use exasol config get, exasol config set, and exasol config reset to inspect or change parameters for the existing presets without deleting local state. To switch presets in the same deployment directory, run exasol destroy --remove before initializing again, or run exasol remove if the deployment resources are already gone.

Runtime tools such as OpenTofu are downloaded on demand and reused from a per-user runtime artifact cache. Use exasol cache list to inspect cached artifacts, exasol cache clean to remove stale artifacts, exasol cache clean --invalid to remove artifacts that fail integrity checks, exasol cache clean --partial-downloads to remove staged partial downloads, exasol cache clean --all to wipe cached artifacts, and exasol diag cache to inspect cache health without changing it. Add --dry-run to a cleanup command to preview what would be removed.

📊 Load Sample Data

To get started quickly, Exasol provides a sample dataset hosted on S3 that you can import using SQL.

You can load it directly by executing this command in your deployment directory (e.g. ~/.exasol/personal/deployments/default for a default deployment):

exasol connect -f sample.sql

Alternatively, connect with a SQL client of your choice and paste the statements below:

CREATE OR REPLACE TABLE PRODUCTS AS (
    IMPORT FROM PARQUET
    AT 'https://exasol-easy-data-access.s3.eu-central-1.amazonaws.com/sample-data/'
    FILE 'online_products.parquet'
);

CREATE OR REPLACE TABLE PRODUCT_REVIEWS AS (
    IMPORT FROM PARQUET
    AT 'https://exasol-easy-data-access.s3.eu-central-1.amazonaws.com/sample-data/'
    FILE 'product_reviews.parquet'
);

Exasol infers the table schema directly from the Parquet files, so there's no need to define columns up front.

| Table | Rows | Size | |---|---|---| | PRODUCTS | 1,000,000 | 27.3 MiB | | PRODUCT_REVIEWS | 1,822,007 | 154.5 MiB |

⏯️ Start and stop Exasol Personal

To save costs, you can temporarily stop Exasol Personal by using the following command:

exasol stop

This stops the compute instance(s) that Exasol Personal is running on.

Networking and the data volumes where the database data is stored will continue to incur costs when compute instances are stopped.

To start Exasol Personal again, use the following command:

exasol start

The IP addresses of the nodes will change when you restart Exasol Personal. Check the output of the start command to know how to connect to the deployment after a restart.

For local deployments, which currently require macOS with at least 8 GB RAM, the launcher manages a local VM runtime and an internal deployment share inside the deployment directory. If you do not configure local VM memory explicitly, it defaults to about 50% of detected host memory. Configured local VM memory must be at least 4096 MB. The initial local database credentials are sys / exasol. exasol shell host opens the local VM shell, and exasol shell container opens a shell inside the local database container.

🗑️ Remove Exasol Personal

To completely remove an Exasol Personal deployment, use exasol destroy. This command deletes the deployment resources and all associated data.

To learn more about this command, use exasol destroy --help.

By default, exasol destroy keeps the local deployment files so you can inspect the deployment or recreate the same preset. To also remove the local deployment directory after deployment resources have been destroyed, run:

exasol destroy --remove

If deployment resources were already deleted manually or you no longer have access to destroy them through the launcher, use the local recovery command:

exasol remove

This removes the local deployment directory. It does not destroy deployment resources.

Deleting the deployment directory and the Exasol Launcher will not remove the resources that were created in the target environment. To completely remove a deployment, you must use the exasol destroy command before deleting the deployment directory.

If you have already deleted the deployment directory and the Exasol Launcher, you must remove the resources manually in the target environment.

For local deployments, exasol destroy deletes the local VM disk/data and launcher-managed share for that deployment.

⚙️ Cloud: Choosing cluster size and compute instance types

By default the launcher deploys a single-node cluster on a memory-optimized instance in the cloud (e.g. r6i.xlarge on AWS, StandardE4sv3 on Azure, standard.extra-large on Exoscale, m2i.4 on STACKIT). To change the number of nodes or the instance type, use the --cluster-size and --instance-type options:

exasol install <preset> --cluster-size <number> --instance-type <string>

If the deployment process is interrupted, resources that were already created will not be removed automatically and cloud resources may continue to accrue cost. In that case, use exasol destroy to clean up the deployment, or remove the resources manually in the target environment.

🔜 Next steps

Once the deployment process is complete, use exasol info for information about how to connect to your Exasol database. The credentials for connecting to the database from a client are stored in the file secrets.json in the deployment directory.

You can also use the built-in SQL client in Exasol Launcher to connect directly to the database from the command line:

exasol connect

To run SQL without entering the interactive shell, pass it directly. Both flags accept multiple ;-separated statements and exit when done, which is convenient for scripting and automation:

exasol connect -c "SELECT 1; SELECT 2"   # run inline statement(s) exasol connect -f script.sql             # run statements from a file

--command and --file are mutually exclusive. In this non-interactive mode, execution stops at the first failing statement and connect exits with a non-zero status so scripts can detect errors. Combine with --json for machine-readable output, or use --csv for CSV output:

exasol connect --csv -c "SELECT * FROM PRODUCTS" > products.csv

With --json, non-interactive execution writes exactly one JSON document to stdout for the full invocation, including multi-statement runs and SQL errors. Interactive exasol connect --json continues to emit one JSON document per executed statement.

In an interactive session, query output is capped at 100 rows by default so a large SELECT doesn't flood the terminal; a note is printed when output is truncated. Piped or --command/--file (non-interactive) execution returns the full result set. Use --max-rows N to set the cap explicitly, or --max-rows 0 for unlimited:

exasol connect --max-rows 0        # return all rows, even interactively echo "SELECT * FROM PRODUCTS;" | exasol connect --max-rows 1000

See also...

• To learn more about how you can connect to your Exasol database and start loading data using the many supported tools and integrations, see Connect to Exasol and Load Data.
• To learn how to use the SQL statements, data types, functions, and other SQL language elements that are supported in Exasol, see SQL reference.

🧑‍💻 Exasol Admin

Exasol Admin is an easy-to-use web interface that you can use to administer your new Exasol database. Instructions on how to access Exasol Admin are shown in the terminal output at the end of the install process.

• To find the Exasol Admin URL after the installation has completed, use exasol info.
• The credentials for connecting to Exasol Admin are stored in the file secrets.json in the deployment directory.

Your browser may show a security warning when connecting to Exasol Admin because of the self-signed certificate. Accept this warning and continue.

Currently, Exasol Admin is only available on cloud deployments.

🔒 Connect using SSH

To connect with SSH to your deployment use one of the following commands:

# Connect to the compute instance your database is running on:
exasol shell host

# Connect to the COS container your node is running on:
exasol shell container

📦 Presets

Exasol Personal uses presets — self-contained directories of templates and config files — to provision infrastructure and install Exasol. Each deployment combines two presets:

• Infrastructure preset — provisions deployment resources for a cloud provider or local system. Built-in: aws, azure, exoscale, stackit, local.
• Installation preset — installs and configures Exasol on the provisioned nodes. Built-in: ubuntu (used by default).

exasol install <infra-preset> [install-preset]
exasol install aws                                       # built-in preset by name exasol install ./my-preset                               # local preset by path (starts with . / ~ or contains /) exasol install https://github.com/org/preset.git         # git repository (HEAD) exasol install https://github.com/org/preset.git@v1.0    # git repository at branch or tag exasol install https://example.com/preset.tar.gz         # remote archive (re-fetched on every run) exasol install file:///path/to/preset-dir                # local directory via URI exasol install file:///path/to/preset.tar.gz             # local archive (re-extracted on every run)

List all available built-in presets, including cloud and local targets:

exasol presets

Local presets

You can store your own preset directories anywhere on your filesystem and pass the path directly to exasol install. This lets you target additional infrastructure platforms or customize provisioning without modifying the launcher.

External presets

In addition to built-in names and local paths, exasol install accepts external sources:

• Git repository — https://github.com/org/preset.git or git@github.com:org/preset.git. Append @branch-or-tag to pin a specific ref. The repository is cloned once per commit and cached; repeated runs at the same commit reuse the cache.
• Remote archive — An https:// or http:// URL ending in .tar.gz, .tgz, or .zip. The archive is re-downloaded on every run because no checksum is provided.
• Local directory via URI — file:///absolute/path. The directory is used directly without copying.
• Local archive via URI — file:///absolute/path/to/preset.tar.gz. The archive is re-extracted on every run.

See doc/presets.md for the full preset contract, caching behavior, and troubleshooting.

Building your own preset

See doc/presets.md for the full preset contract: manifest schema, required output artifacts, variable channels, and the reference implementations in assets/infrastructure.

🖥️ System Requirements

The Exasol Launcher runs on:

• macOS — 12 (Monterey) or later, on Apple Silicon or Intel.
• Linux — amd64 or arm64.
• Windows — 10 or later, amd64.

Where the database runs depends on the deployment type:

• Cloud deployments — the database runs on your provider's infrastructure, so any supported launcher platform above works. The launcher provisions Ubuntu 22.04 LTS (x86-64) compute instances on all cloud providers.
• Local deployments — the database runs in a VM on your machine, which requires macOS 15 (Sequoia) or later on Apple Silicon, with at least 8 GB RAM.

🚧 Limitations

Local deployments are intended for development and exploration and do not yet support every feature of a cloud deployment:

• UDFs — user-defined functions are not available yet (coming soon).
• Virtual schemas — virtual schemas are not available yet (coming soon).
• Admin UI — the Administration UI is not available yet (coming soon).
• Multi-node clusters — a local deployment is a single VM on your machine by design and always runs as a single node.

Cloud deployments support all of the above.

⚖️ Licensing

The Exasol Launcher source code in this repository is open-source software licensed under the MIT License. You are free to use, modify, and distribute it. Contributions are made under the same terms.

The launcher installs the Exasol Database, which is proprietary software provided by Exasol AG, free for personal use. By deploying it with exasol install, you accept the Exasol Personal End User License Agreement (EULA).

📚 Resources & Documentation

Get the most out of Exasol with these resources:

• Exasol Documentation — Complete database documentation
• Connect to Exasol — Driver downloads and client setup
• Load Data — Import data into your database
• SQL Reference — Complete SQL syntax reference
• Exasol Community — Ask questions and share knowledge (use tag exasol-personal)