Skip to main content

Installation

There are two options depending on whether you want to install Python or not:

Option 1: Ready-to-use executable​

The executable is a self-contained zip archive. No Python, no pip, no virtual environment needed.

1. Download the zip

Either use the zip you received directly, or download it from the GitHub releases page — expand Assets and pick the file for your platform (e.g. testbench-ai-service-x.y.z-win_amd64.zip).

2. Extract to a permanent location

Unzip to a folder you won't move later, for example C:\TestBenchAIService\:

C:\TestBenchAIService\
  testbench-ai-service.exe
  <support files>

3. Verify installation from the command line

C:\TestBenchAIService\testbench-ai-service.exe --version
Running as a Windows service

Point NSSM or FireDaemon directly to testbench-ai-service.exe in the extracted folder. See the Windows service guide.

Option 2: Install with Python​

Use this option if you have Python installed or are about to install it.

Requirements​

  • Python 3.10–3.14 — 3.13 recommended (download)
  • pip (included with Python)

Set up a virtual environment​

note

If you are installing from source: create the virtual environment inside the cloned repository instead.

A virtual environment keeps the installation isolated and provides a stable, predictable path to the executable. Navigate to the directory where you want the service to live, then run:

python -m venv .venv
# Windows
.venv\Scripts\activate

# Linux / macOS
source .venv/bin/activate
Running as a Windows service

With a venv, point NSSM or FireDaemon to .venv\Scripts\testbench-ai-service.exe (full absolute path). See the Windows service guide.

pip install testbench-ai-service

Verify installation:

testbench-ai-service --version

From a wheel file (offline)​

Use this if you received a .whl file or downloaded one from the GitHub releases page (look for testbench_ai_service-x.y.z-py3-none-any.whl in the Assets).

pip install testbench_ai_service-x.y.z-py3-none-any.whl

Verify installation:

testbench-ai-service --version
Fully offline install

By default pip still fetches dependencies from PyPI. To install on a machine with no internet access:

On another machine which has internet access, download the wheel and all its dependencies into a local folder:

pip download "testbench_ai_service-x.y.z-py3-none-any.whl" -d ./wheels

This saves every required wheel into ./wheels/. Copy that folder together with the .whl file to the target machine.

On the target machine, install entirely from the local folder:

pip install --no-index --find-links ./wheels testbench_ai_service-x.y.z-py3-none-any.whl

From source (developers)​

git clone https://github.com/imbus/testbench-ai-service.git
cd testbench-ai-service

Create and activate the virtual environment inside the cloned repository. See Set up a virtual environment.

Install with development dependencies:

pip install -e .[dev]

Verify installation:

testbench-ai-service --version

Next steps​

Head to the Quickstart to configure and start the service.