Installation

Requirements

• Python 3.10 or newer

• Tesseract OCR (optional — used as a fallback when PaddleOCR times out)

Note

No Ollama required. LLM models are downloaded automatically from HuggingFace on first use (~4 GB per model) and cached in ~/.cache/huggingface/hub. On Apple Silicon the mlx-lm backend is used; on other platforms transformers is used.

Install from PyPI

pip install finamt

For CLI usage, installing via pipx is recommended. It places finamt into its own isolated virtual environment so its dependencies never conflict with other projects, while still making the finamt command available system-wide without activating a virtualenv:

pipx install finamt

Optional extras

Install development tools:

pip install "finamt[dev]"

Install documentation tools:

pip install "finamt[docs]"

Installing Tesseract (optional)

# macOS
brew install tesseract tesseract-lang

# Ubuntu / Debian
sudo apt install tesseract-ocr tesseract-ocr-deu

Tesseract is only invoked if PaddleOCR fails or exceeds its timeout, so it is not required for most use cases.