Skip to content

Install Matemium

Matemium is a desktop application with a bundled local rendering engine. Native release automation exists for four launch targets:

Linux x86_64.deb and .AppImage
Windows x86_64NSIS .exe and .msi
macOS Apple SiliconNative .dmg
macOS IntelNative .dmg
  1. Open the Matemium download page.
  2. Select an available build for your operating system and processor.
  3. Install the application using your platform’s normal installer flow.
  4. Start Matemium and allow the bundled engine to initialize.
  5. Download a local language model only if you want offline AI assistance.

The installer contains the Matemium UI, Rust shell, and native Python sidecar. It does not contain FFmpeg, a LaTeX distribution, local language models, or AI provider keys.

  • Linux: Ubuntu 24.04 or compatible on x86_64, FFmpeg, and TeX Live
  • Windows: Windows 10/11 on x86_64, FFmpeg, and MiKTeX
  • macOS: macOS 12 or newer, FFmpeg, and MacTeX or BasicTeX
  • Enough free disk space for project media, render caches, output, and any optional local model

If a .deb reports an unresolved FFmpeg dependency:

Terminal window
sudo apt -f install

Matemium can render without a cloud AI provider. To use external chat or agent workflows:

  1. Open Settings.
  2. Choose an OpenAI-compatible provider.
  3. Connect your account or add your own provider key.
  4. Select a model for the current workflow.

Provider billing and limits belong to the provider you select. Matemium does not resell model access.

Use this path for engine development or contribution—not as the normal desktop installation.

Terminal window
git clone https://github.com/fargonee/matemium.git
cd matemium
python -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"
./matemium.sh demo

Source renders and packaged desktop renders both require FFmpeg and a usable LaTeX installation on the host.

Terminal window
./matemium.sh list
./matemium.sh demo

A successful demo writes output beneath outputs/demo/.

Create your first project →