jevgram

jevgram is a small command-line tool that estimates whether a document was primarily written by AI or by a human using JEV. It is intended as a lightweight, scriptable alternative for occasional authorship checks.

Supported inputs

• Plain text and source-like text files, such as .txt, .md, .tex, and similar files
• PDF files
• DOCX files

PDF extraction uses pypdf. DOCX extraction is handled directly from the Word XML contained in the .docx file.

Install

With pipx:

pipx install jevgram

With uvx:

uvx jevgram --help

API key

jevgram calls the TypeSafe JEV API. Set one of these environment variables before running it:

export TYPESAFE_API_KEY="your-key"
# or
export JEV_API_KEY="your-key"

You can also pass a key directly with --api-key, but environment variables are safer for shell history and scripts.

Usage

jevgram document.txt
jevgram paper.pdf
jevgram manuscript.docx --json

By default, long files are sampled down to 32,000 characters using the beginning, middle, and end of the extracted text. To send the whole extracted document, use:

jevgram document.txt --max-chars 0

To inspect extraction without calling JEV:

jevgram document.pdf --extract-only

To inspect the request body without calling JEV:

jevgram document.txt --dump-request

Output

Human-readable output includes the predicted class, AI probability, human probability, confidence when provided by JEV, and whether the input was truncated. --json produces machine-readable output for scripts.

Notes

Authorship detection is probabilistic. Treat the result as one signal, not as proof. Results can vary with document length, extraction quality, formatting, and how much of the document is sent to JEV.