Skip to content
tomo

Installation

Install tomo from Go, a release archive, a Linux package, Homebrew or Scoop, or the container image, and what it needs to run.

tomo is a single static binary. It is built CGO_ENABLED=0, so it links nothing and runs the same on any machine of its platform. It needs a provider API key to talk to a model (for example ANTHROPIC_API_KEY), but it has no login and no account of its own: there is no tomo service to sign in to.

Go

go install github.com/tamnd/tomo/cmd/tomo@latest

Release archives and Linux packages

Every release attaches tar.gz archives (and a .zip for Windows) for Linux, macOS, Windows, and FreeBSD, plus .deb, .rpm, and .apk packages and a checksums.txt. Download the one for your platform, extract tomo, and put it on your PATH.

# Debian/Ubuntu
sudo dpkg -i tomo_*_linux_amd64.deb

# Fedora/RHEL
sudo rpm -i tomo_*_linux_amd64.rpm

Homebrew and Scoop

Homebrew and Scoop manifests publish alongside each release once their taps are configured, so a brew install and a scoop install land you the same binary on macOS and Windows.

Container

The image carries the binary and nothing else. tomo keeps its config and data under $HOME/.tomo, so mount a volume at the container's home and everything (the config, the sqlite ledger, memory, and the audit log) persists across runs:

docker run --rm -it \
  -e HOME=/data -v "$PWD/tomo-data:/data" \
  -e ANTHROPIC_API_KEY \
  -p 8765:8765 \
  ghcr.io/tamnd/tomo serve --addr 0.0.0.0:8765

The volume at /data holds ~/.tomo, -p 8765:8765 publishes the web chat, and binding --addr 0.0.0.0:8765 makes it reachable from outside the container. Run onboard once against the same volume first to write the starter config.

What it needs to run

  • A config at ~/.tomo/config.yaml. Run tomo onboard to write a starter one (see the quick start).
  • A provider key in your environment, referenced from the config with ${VAR}. With the default config that is ANTHROPIC_API_KEY; point a provider at a local server instead and it can be whatever that server wants, or nothing.

Optional external tools for voice

Voice is off until you configure it, and it leans on external binaries that tomo discovers on your PATH and never bundles:

  • whisper-cli (whisper.cpp) transcribes inbound voice notes.
  • piper speaks replies back.
  • ffmpeg decodes inbound clips and encodes the spoken reply.

All of it runs locally, so no audio leaves the machine. See the voice guide for the setup.

Next: the quick start.

2 min read