Get started

Install Livemore Bio

Choose a path for your computer. The install itself takes several minutes, mostly for the verification suite, and a full evaluation takes about 15 minutes. You do not need a GitHub account, git or a login.

Your computer Use
Windows Docker
macOS Native install or Docker
Linux Native install or Docker

Windows is not supported natively, on purpose. Livemore Bio protects local health-like data with owner-only POSIX file permissions, which Windows cannot express. Rather than pretend, it refuses to run natively and points you to Docker, which gives you the identical product.

Requirements

Tool Needed for Required?
Python 3.11 everything required
curl downloading the release required (preinstalled on macOS and most Linux)
C compiler + SUNDIALS the cardiac electrophysiology engine (Myokit compiles CVODE code at run time) recommended
Node.js 20+ console diagnostics, the anatomy and atlas viewer checks, and about 950 of the verification tests (they are skipped without it) recommended
Playwright 1.58.2 browser-driven checks of the console. The verification suite pins this exact version recommended
Docker Desktop the container path (required on Windows) optional

Without the C compiler and SUNDIALS, Livemore Bio still installs and runs. The cardiac engine reports CARDIAC_TOOLCHAIN_UNAVAILABLE, names what to install, and its tests skip rather than failing the installation. If Node.js or Playwright is missing, the checks that need them are skipped and Livemore Bio tells you so. livemore doctor reports which tools it found.

macOS

# the C compiler comes with the Xcode command line tools
xcode-select --install 2>/dev/null || true
brew install python@3.11 node sundials
npm install -g playwright@1.58.2 && npx playwright install chromium

Ubuntu and Debian

On Ubuntu 24.04, python3.11 comes from the deadsnakes PPA.

sudo apt update && sudo apt install -y python3.11 python3.11-venv curl build-essential libsundials-dev
curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash - && sudo apt install -y nodejs
sudo npm install -g playwright@1.58.2 && npx playwright install chromium

Fedora and RHEL

sudo dnf install -y python3.11 gcc gcc-c++ make sundials-devel nodejs

Native install (macOS, Linux, WSL2)

curl -fsSL https://raw.githubusercontent.com/MannyAmah/livemorebio-dist/main/install.sh | bash

The installer:

  • downloads the published release over HTTPS and checks its SHA-256 before unpacking anything;
  • builds its own isolated Python environment from a hash-locked dependency set;
  • downloads the pinned scientific reference models (Human-GEM 2.0);
  • runs the complete test suite against a throwaway data directory;
  • activates the livemore command only if everything passed.

If any step fails, it stops, names the step and leaves your previous install untouched. It never half-installs. Expect several minutes, mostly for the verification suite.

Pin the exact build

If you were sent a checksum, pin it. The install then refuses anything that is not bit-for-bit the published archive:

curl -fsSL https://raw.githubusercontent.com/MannyAmah/livemorebio-dist/main/install.sh \
  | LIVEMORE_SHA256=<the checksum we sent you> bash

The current channel manifest is stable.json. Get the checksum you pin from us through a separate channel (your contract, an email or the data room), not from the distribution repository.

If livemore is not found afterwards, open a new terminal or add ~/.local/bin to your PATH.

Docker (any operating system)

Docker runs the same three services in containers. You need Docker Desktop and nothing else.

curl -fsSL https://raw.githubusercontent.com/MannyAmah/livemorebio-dist/main/livemorebio-stable.tar.gz | tar -xz
cd livemorebio-*/

# three secrets that stay on your machine; never commit or email them
export LIVEMORE_API_TOKEN=$(openssl rand -hex 32)
export LIVEMORE_PATCH_STORAGE_KEY=$(openssl rand -base64 32 | tr '+/' '-_')
export LIVEMORE_PATCH_ADMISSION_HMAC_KEY=$(openssl rand -base64 32 | tr '+/' '-_')

docker compose up -d
curl -fsS http://127.0.0.1:8850/health      # then open http://127.0.0.1:8850

On Windows PowerShell:

curl.exe -fsSLO https://raw.githubusercontent.com/MannyAmah/livemorebio-dist/main/livemorebio-stable.tar.gz
tar -xzf livemorebio-stable.tar.gz
cd (Get-ChildItem -Directory livemorebio-*)[0]

$env:LIVEMORE_API_TOKEN = -join ((1..32) | ForEach-Object { '{0:x2}' -f (Get-Random -Max 256) })
$env:LIVEMORE_PATCH_STORAGE_KEY = [Convert]::ToBase64String((1..32 | ForEach-Object { Get-Random -Max 256 })).Replace('+','-').Replace('/','_')
$env:LIVEMORE_PATCH_ADMISSION_HMAC_KEY = [Convert]::ToBase64String((1..32 | ForEach-Object { Get-Random -Max 256 })).Replace('+','-').Replace('/','_')

docker compose up -d

The first run downloads and builds; later runs take seconds. To stop, run docker compose down. Add -v to also erase stored data.

How the image is built

The image is built from the same hash-locked dependency set as the native installer, and the build runs the full test suite. If a scientific kernel cannot compile or a test fails, the build fails instead of producing an image without its engines. The image does not pre-install the Human-GEM reference model.

Manual install from the release archive

For contributors who want an editable environment instead of the managed installer. This path skips the installer's test gate, so run the suite yourself.

curl -fsSL https://raw.githubusercontent.com/MannyAmah/livemorebio-dist/main/livemorebio-stable.tar.gz | tar -xz
cd livemorebio-*/
python3.11 -m venv .venv
./.venv/bin/python -m pip install --require-hashes -r requirements-python311.lock
./.venv/bin/python -m pip install --no-deps --no-build-isolation -e .
./.venv/bin/livemore models install human-gem
LIVEMORE_TEST_GENOME_MODELS=1 ./.venv/bin/python -m pytest -q   # also runs the Human-GEM integration tests, as the installer does

The Docker and manual paths download the archive with curl | tar, so check it against livemorebio-stable.tar.gz.sha256 yourself.

Check the install

livemore doctor

You want the last line to read READY. The check covers Python packages, the reference models, the data directory's permissions, your keys and each service. It exits non-zero if anything required is missing, so it is safe to use in scripts.

Next: run the quickstart.

Questions about this page?

Email us