Get started¶

Your first Pipe needs Python and a few local rows. No database server or Docker setup is required.

Install¶

Use Python 3.11 or newer in your usual virtual environment:

python -m pip install --upgrade meerschaum
mrsm --version

The package includes the native core and the mrsm command. python -m meerschaum works too.

Using DataFrames in Python?

New installations return Polars DataFrames. Install the matching extra:

python -m pip install 'meerschaum[polars]'

For pandas, install meerschaum[pandas] and choose dataframe: pandas in meerschaum.yaml. Existing roots from v4 keep pandas. You can also read rows as plain dictionaries with pipe.get_data(as_docs=True).

See Reading data.

Make a workspace¶

Keep the tutorial's configuration and SQLite database together:

Linux / macOSPowerShell

mkdir mrsm-demo
cd mrsm-demo
mkdir .meerschaum
export MRSM_ROOT_DIR="$PWD/.meerschaum"

mkdir mrsm-demo
cd mrsm-demo
mkdir .meerschaum
$env:MRSM_ROOT_DIR = "$PWD/.meerschaum"

sql:local is a built-in SQLite connector. It stores your data inside this workspace:

mrsm show pipes -i sql:local

You should see No pipes to show. Keep this terminal open for the next steps.

Build your first Pipe →

Other ways to run Meerschaum

Your database: add a SQL connector. The stack supplies a ready-to-use TimescaleDB database through Docker.

Standalone command: cargo install meerschaum --locked builds the stable release's native CLI. Prebuilt binaries are listed on GitHub Releases.

Docker: see the image variants.

macOS binaries: downloaded binaries are unsigned. If Gatekeeper blocks one, remove its quarantine flag with xattr -d com.apple.quarantine ./mrsm-<version>-<target>, then run chmod +x on the file. The pip installation does not need this step.

Windows: run commands in the foreground. Background jobs need Linux, macOS, Docker, or WSL.

Upgrading an existing installation? Read Migrating to v5 first.