upspawn-ocr-cli 0.1.0b4

Modern CLI to extract text from PDFs using Mistral cloud or local Ollama models (glm-ocr, deepseek-ocr, LightOnOCR-2).

Mistral OCR CLI

Modern, polished CLI to extract text from PDFs using the Mistral OCR API.

Features

• Elegant TUI with progress bars and rich output
• Single file or batch processing
• Output in text, JSON, or Markdown
• Parallel batch processing with --jobs
• Config helper and .env support

Quickstart

1. Install

uv tool install mistral-ocr-cli  # via pipx-like tool install
# or
uv pip install mistral-ocr-cli   # into current environment

2. Configure API key

export MISTRAL_API_KEY=your_key_here
# or
echo "MISTRAL_API_KEY=your_key_here" >> .env

3. Extract text

ocr extract file.pdf -o out.txt
ocr extract file1.pdf file2.pdf --batch --output-dir outputs --jobs 4

Usage

ocr extract [OPTIONS] FILES...

Options:
  -o, --output PATH            Output file (single-file mode)
  -f, --format [text|json|markdown]
  -b, --batch                  Enable batch mode
  -O, --output-dir PATH        Directory for batch outputs
  -j, --jobs INTEGER RANGE     Parallel jobs for batch [default: 1]
  -v, --verbose                Verbose logs
  -q, --quiet                  Only errors
  --version                    Show version
  --help                       Show help

Programmatic use

from ocr.pdf2text import pdf_to_text

text = pdf_to_text("/path/file.pdf")

Development

uv pip install -e .[dev]
uv run pre-commit install
uv run pytest -q

Releasing is handled via standard tags and GitHub Releases.

License

MIT

Test coverage

# Terminal report
make coverage

# HTML report in htmlcov/
make coverhtml