Downloads

The Spark runtime is an AI-first language runtime: ask, classify, extract, pipeline, review, model analyze, and voice — one reviewable .spark file per workflow. Dry-run offline with fixtures; go live with any standard OpenAI-compatible API and OPENAI_API_KEY.

Graphical installers (recommended)

Primary downloads are setup wizards — not curl|bash one-liners. CLI scripts remain available as fallbacks.

Platform	Graphical installer	How to launch
Linux x86_64 / arm64	`.deb` (Software Center / GNOME Software) or `spark-install-linux-gui.sh` (zenity progress wizard)	Double-click the `.deb`, or run the GUI script; after install, find Install Spark in Applications (zenity helper).
macOS	`.pkg` with welcome + conclusion pages (Apple Silicon or Intel build)	Double-click the `.pkg` in Finder — macOS Installer.app walks through the GUI steps.
Windows	`spark-runtime-windows-*-setup.exe` (Inno Setup wizard: install folder, PATH checkbox, first-run build)	Double-click the `.exe`. If only a `.zip` is listed, extract and run `Setup.bat` (CLI fallback).

OS	CPU	Kind	Installer (recommended)	Also available
Loading packages from manifest…

manifest.json

AI quick start — Linux

Graphical: download the .deb and open it in Software Center, or run bash spark-install-linux-gui.sh (needs zenity).

After install:

# .deb (root)
sudo dpkg -i spark-runtime-linux-x86_64-0.6.0.deb
spark --dry-run /opt/spark/examples/hello.spark
spark --dry-run /opt/spark/examples/classify_intent.spark

# curl|bash (user install, no root)
curl -fsSL https://sparklang.dev/downloads/spark-install-linux.sh | bash -s -- --user
spark --dry-run ~/.local/spark/examples/hello.spark

Live ask: set OPENAI_API_KEY, then spark --live /opt/spark/examples/ask_live.spark. See Program with Spark §3.

AI quick start — macOS

Graphical: double-click the .pkg for your CPU (arm64 Apple Silicon or x86_64 Intel). CLI fallback:

curl -fsSL https://sparklang.dev/downloads/spark-install-macos.sh | bash

/usr/local/spark/bin/spark --dry-run /usr/local/spark/examples/hello.spark

The graphical installer sets up Spark on your Mac. For full companion binaries, use the Linux package or build from source.

AI quick start — Windows

Graphical: run spark-runtime-windows-x86_64-*-setup.exe (Inno Setup wizard). Choose install folder, optional PATH, then allow the post-install first-run build.

bin\spark --dry-run examples\hello.spark

Windows ships the Spark runtime. For spark-ask-http, speech, and gateway companions, use WSL with the Linux x86_64 installer.

Verify

curl -sO https://sparklang.dev/downloads/manifest.json
curl -sO https://sparklang.dev/downloads/spark-runtime-linux-x86_64-0.6.0.deb
sha256sum -c <(python3 -c "
import json; m=json.load(open('manifest.json'))
p=next(x for x in m['platforms'] if x.get('primary') and x['id']=='linux-x86_64')
print(p['sha256']+'  '+p['file'])
")

Build from source

Download the runtime kit for your platform from the table above, extract the archive, then run make in the extracted directory. Packages here match release 0.6.0. Need a different build or platform? Use the contact link on this page.