Install¶

Install script (Linux / macOS)¶

The fastest way to get the standalone datapress CLI without a Rust toolchain. It downloads the prebuilt binary (both backends bundled) for your platform, verifies its checksum, and installs it into ~/.local/bin — no sudo, and your shell profile is never edited:

curl -LsSf https://datap-rs.org/install.sh | sh

If ~/.local/bin is not already on your PATH, the script prints the exact line to add. Override the target directory or version with:

# Install a specific version into a custom directory.
DATAPRESS_INSTALL_DIR="$HOME/bin" DATAPRESS_VERSION=0.4.4 \
  sh -c "$(curl -LsSf https://datap-rs.org/install.sh)"

Install script (Windows)¶

powershell -ExecutionPolicy ByPass -c "irm https://datap-rs.org/install.ps1 | iex"

This installs into %LOCALAPPDATA%\datapress\bin and adds it to your user PATH. Open a new terminal afterwards.

Homebrew (macOS / Linux)¶

brew install jeroenflvr/tap/datapress

Apple Silicon macOS and Linux (x86_64 / aarch64) are covered by prebuilt bottles. On Intel Macs, use the crates.io install below.

winget (Windows)¶

winget install datap-rs.DataPress

Docker¶

The image installs the prebuilt Linux binary onto a minimal distroless/cc base and is multi-arch (linux/amd64 + linux/arm64). Mount a datasets.toml at /etc/datapress/datasets.toml — it must set listen = "0.0.0.0" for the published port to be reachable:

docker run --rm -p 8080:8080 \
  -v "$PWD/datasets.toml:/etc/datapress/datasets.toml:ro" \
  jeroenflvr/datapress:latest

Tags: :<version>, :<major.minor>, and :latest.

Prebuilt binary (crates.io)¶

The quickest way to get a server without cloning the repo. The unified datapress binary bundles both the DuckDB and DataFusion backends and selects one at runtime from server.backend in your datasets.toml:

cargo install datapress
datapress                  # reads ./datasets.toml (or $DATASETS_CONFIG)

For a slimmer single-backend build, or to opt into the docs / Swagger / metrics / auth features:

cargo install datapress --no-default-features --features duckdb
cargo install datapress --features swagger,auth,metrics

Configuration discovery¶

The installed datapress binary finds its config in this order (first match wins):

--config <FILE>
$DATAPRESS_CONFIG_FILE
./datasets.toml
$HOME/datasets.toml

Generate a commented starter template with datapress init (writes to $HOME when no directory is given):

datapress init                 # ~/datasets.toml.template
datapress init ./config        # ./config/datasets.toml.template
cp ~/datasets.toml.template ~/datasets.toml   # then edit and run `datapress`

From source (Rust binaries)¶

Two binaries live in the workspace, one per backend. Both build from a checkout of the repo:

git clone https://github.com/jeroenflvr/datapress.git
cd datapress

cargo build --release -p datapress-duckdb
cargo build --release -p datapress-datafusion

The release binaries land in target/release/datapress-{duckdb,datafusion}.

If you have task installed:

task build:duckdb          # DuckDB binary
task build:datafusion      # DataFusion binary
task build                 # both

Python wheel¶

The Python wheel datap-rs bundles both engines and lets you pick one at runtime via DataPressConfig(backend=...).

pip install datap-rs
# or
uv pip install datap-rs

Wheels are published for Linux (x86_64/aarch64), macOS (arm64), and Windows (x86_64) against CPython 3.9+ (abi3).

Building the wheel from source:

task py:develop     # editable install into ./.venv (uses uv + maturin)
task py:build       # release wheel into ./target/wheels/

Optional features¶

Feature	Crate	Purpose
`docs`	`datapress-core`	Embed this documentation site into the binary. Disabled by default.

Enable at build time:

task docs:build
cargo build --release -p datapress-duckdb --features docs

The same features are also forwarded by the unified datapress crate, so they work with cargo install datapress --features docs too.

See Configuration › Documentation site for the runtime switch.