Installation

FlowRoute requires Python 3.10 or newer.

Core router

Create an isolated environment and install the published package:

python -m venv .venv
source .venv/bin/activate
pip install flowroute

On Windows PowerShell, activate the environment with:

.venv\Scripts\Activate.ps1

The core install provides:

• the flowroute Python package;
• the flowroute route and flowroute evaluate commands;
• the local TF-IDF retriever and lexical verifier; and
• YAML catalog validation.

It does not include the demo catalog, fixtures, or training scripts. Those live in the repository; see From source if you need them.

Optional extras

HTTP APIHugging FaceModel trainingDevelopmentDocumentation

pip install "flowroute[api]"

Adds FastAPI and Uvicorn for flowroute serve.

pip install "flowroute[hf]"

Adds the learned inference backends. This install is significantly larger because it includes Torch.

pip install "flowroute[training]"

Adds Datasets and Accelerate to the learned-backend dependencies. The training scripts themselves ship only in the repository.

pip install -e ".[dev]"

Adds pytest, Ruff, mypy, build, and HTTPX. Run this from a clone, not against the published package.

pip install -r requirements-docs.txt

Adds MkDocs and Material for MkDocs. Requires a clone.

From source

Work from a clone when you need the demo catalog, the fixture suite, the training scripts, or a development install:

git clone https://github.com/open-first/FlowRoute.git
cd FlowRoute
python -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"

An editable install points the flowroute command at your working tree, so source edits take effect without reinstalling.

Verify the install

flowroute --help
python -c "import flowroute; print(flowroute.__version__)"

From a clone, also run the test suite:

pytest -q

Dependency groups

Install	Main dependencies	Intended use
pip install flowroute	NumPy, Pydantic, PyYAML, regex, scikit-learn	Local baseline and library
pip install "flowroute[api]"	FastAPI, Uvicorn	HTTP service
pip install "flowroute[hf]"	Torch, Transformers, Sentence Transformers	Learned inference backends
pip install "flowroute[training]"	Datasets, Accelerate, and learned-backend dependencies	Dataset processing and training
pip install -e ".[dev]"	pytest, Ruff, mypy, build, HTTPX	Development and tests (clone only)

Next: route the first request.