Skip to main content

Zero Slop

Find AI-sounding writing. Keep the source intact.

A prose linter that scores text 0–100 for stock phrasing, mechanical rhythm, vague claims, and canned formatting, then tells you which phrases produced the number. It runs offline, uses only the Python standard library, and has no runtime dependencies.

The score describes the writing. It does not identify who wrote it. Zero Slop is not an AI detector and must not be used as one.

Install

pip install zero-slop

Use it

slopscore draft.md                      # readable report
slopscore --json draft.md               # machine-readable
slopscore --explain draft.md            # report, reasons, line-by-line map
slopscore --batch docs/ --gate 25       # exits non-zero above the threshold
slopscore --fidelity before.md after.md # did the rewrite keep the facts?
cat draft.md | slopscore                # stdin

As a library

from zero_slop import load_patterns, score_text

data = load_patterns()
score_text("We are thrilled to announce a transformative solution.", data)["ai_likelihood"]
# 99.3

In CI

- uses: manavmishra/ZeroSlop@v2.11.4
  with:
    path: docs
    gate: "25"

As a pre-commit hook

repos:
  - repo: https://github.com/manavmishra/ZeroSlop
    rev: v2.11.4
    hooks:
      - id: zero-slop
        args: [--gate, "25"]

What the score means

Human reference samples score 9–21. Unedited model output averages 77. Those are reference points, not boundaries. The scoring was run over 7,627 abstracts from the MIT-licensed RAID+ dataset and all 2,187 Beemo records; the drafts, outputs, and hashes are in the repository so the numbers can be reproduced.

The scorer handles about 500 documents per second on a single machine, so it is usable as a batch gate rather than only an interactive tool.

Rewriting

This package is the scorer. The full editorial workflow — where your own AI assistant rewrites the draft across copy-desk and read-aloud passes, and a local fact gate rejects any revision that adds or drops a name, number, quotation, or link — ships as an Agent Skill:

npx skills add manavmishra/ZeroSlop --global

Metadata

Release files for zero-slop 2.12.23

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for zero-slop 2.12.23
File Size Uploaded
zero_slop-2.12.23.tar.gz 188.7 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for zero-slop 2.12.23
File Interpreter ABI Platform
zero_slop-2.12.23-py3-none-any.whl Python 3 none any Details

Total release size: 271.3 kB

Release files / zero_slop-2.12.23.tar.gz

Download URL zero_slop-2.12.23.tar.gz
Size 188.7 kB
Tags Source
SHA-256 checksum
How to use checksums
a92a980b08bca95ab22d74f9e1e0d0f9d451b6dc707cc50614e55735a50f2973
BLAKE2b-256 checksum
How to use checksums
f8bd7b49c3b376139da1c7de622396cc022888609cb7d251dc6be68c53d0ccac
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.7

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 14, 2026.

Transparency log

Release files / zero_slop-2.12.23-py3-none-any.whl

Download URL zero_slop-2.12.23-py3-none-any.whl
Size 82.6 kB
Tags Python 3
SHA-256 checksum
How to use checksums
18cbd3706037d96f235c70f7fdbc78ce265ac4be88b7554e62cea67f55212df4
BLAKE2b-256 checksum
How to use checksums
bf6756b6d8968a9c62966f1cab545706040b32593014480ea0d3f74d3ff9eaec
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.7

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 14, 2026.

Transparency log
Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page