The first alpha is out.

Version 0.1.0-alpha.1 is a Mac app for Apple silicon, and a Docker setup for a home server or a team. It’s an alpha, so expect rough edges. Pick the one you’re after.

Use the app

The first alpha is a Mac app for Apple silicon, on macOS 15 or newer. Windows and Linux builds aren’t part of it yet; you can build those from source today.

What’s in the Mac app

FFmpeg, qpdf, Poppler, ExifTool and 7-Zip come inside the app. ImageMagick, Tesseract, Ghostscript, Pandoc, whisper.cpp and Real-ESRGAN download into the app’s own folder the first time a tool needs them, or from Settings → Dependencies, at versions tested with the release. Nothing is installed system-wide. LibreOffice, Piper and rembg aren’t managed yet: install them yourself if you need them.

Opening it the first time

The alpha isn’t signed with an Apple Developer ID yet. If macOS says it can’t verify the developer, click Done, open System Settings → Privacy & Security and click Open Anyway. You only do this once. Intel Macs aren’t supported by this build. Installing the alpha

Build it from source

You need Rust 1.95 or newer, Node.js 20 or newer, a C compiler and the Tauri system packages for your platform. cargo tauri build makes a .dmg on macOS, .msi and .exe on Windows, and .deb, .rpm and AppImage on Linux.

A build from source uses the engines installed on your system. They’re optional, and each one unlocks more tools: dogubako deps lists what’s missing and the command to install it.

Building, in the docs

# build the desktop app
$ git clone https://github.com/Mischa-dev/Dogubako
$ cd Dogubako
$ (cd web && npm ci && npm run build)
$ cargo install tauri-cli --version '^2'
$ (cd apps/desktop/src-tauri && cargo tauri build)

# or just the CLI
$ cargo install --path crates/cli

# which engines are installed, and how to add the rest
$ dogubako deps

Host it on a server

One install for a household or a team, on any Linux machine with Docker. Everyone uses it from a browser, phones included, and agents connect over MCP. Every engine is built into the image.

  1. Get the code

    There’s no public prebuilt image yet, so compose builds it from source.

    $ git clone https://github.com/Mischa-dev/Dogubako
    $ cd Dogubako
  2. Build and start it

    The server runs as user 10001 inside the container, so give it the data folder first. The first build takes a while; the image is about 1.5 GB. Copy .env.example to .env first if you want LibreOffice or a shared model for plain words.

    $ mkdir -p data && sudo chown 10001 data
    $ docker compose up -d --build
  3. Read the setup token

    On first start the server prints a one-time token for creating the owner account. If it restarts before the owner exists, it prints a new one, so use the newest.

    $ docker compose logs dogubako \
        | grep "Setup token" | tail -n 1
  4. Open it and create the owner

    Open http://localhost:8420 on the server, or its Tailscale address from another device, enter the token and pick a username and password. Add people and agents afterwards in Admin → Users & agents.

Keep it on your own network

The compose file only listens on 127.0.0.1. To use it from your phone or laptop, add the machine to your tailnet and run tailscale serve --bg 8420: you get HTTPS on your own devices and nothing is opened to the internet. On a LAN, change the port mapping to "8420:8420". Don’t expose port 8420 directly to the internet; accounts are always required, but there’s no reason to invite the traffic.

Or have your agent set it up.

If a coding agent such as Claude Code runs on the server, give it this prompt. It points to a page written for agents, with each step, how to connect itself over MCP, and what it has to ask you first.

Read agent-install.md

Prompt

Install Dogubako on this server and connect it to me over MCP. Follow the instructions at https://dogubako.app/agent-install.md and ask me before any step that opens ports or changes the firewall.