Metadata-Version: 2.4
Name: ai-text-watermark-audit
Version: 0.1.0
Summary: Local-first auditing for invisible Unicode signals and AI text watermark claims
Author: AI Text Watermark Audit contributors
License: MIT
Project-URL: Homepage, https://github.com/steven-panxd/ai-text-watermark-audit
Project-URL: Documentation, https://github.com/steven-panxd/ai-text-watermark-audit/tree/main/docs
Project-URL: Issues, https://github.com/steven-panxd/ai-text-watermark-audit/issues
Keywords: ai-watermark,claude,llm-watermark,text-forensics,unicode-security
Classifier: Development Status :: 3 - Alpha
Classifier: Environment :: Console
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Security
Classifier: Topic :: Text Processing :: General
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Dynamic: license-file

# AI Text Watermark Audit

[![CI](https://github.com/steven-panxd/ai-text-watermark-audit/actions/workflows/ci.yml/badge.svg)](https://github.com/steven-panxd/ai-text-watermark-audit/actions/workflows/ci.yml)
[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)

Local-first tools for examining invisible signals in text and testing claims about AI text watermarks. The first research target is Claude; the scanner itself is provider-neutral.

[简体中文](README.zh-CN.md) · [Claude watermark notes](docs/claude-watermark.md) · [Methodology](docs/methodology.md) · [FAQ](docs/faq.md)

## Why this exists

Anthropic says supported Claude models embed an imperceptible watermark in generated text. It has not yet published the technical detector required to verify that statistical signal. This project keeps two questions separate:

1. Does a document contain observable Unicode or formatting signals?
2. Does an official detector verify a provider watermark?

Today, `textmark` can answer the first question. For Claude's model-level watermark, it reports the public evidence and an honest `not-publicly-documented` detector status instead of guessing.

## Install

Python 3.10 or newer is required.

```bash
pipx install git+https://github.com/steven-panxd/ai-text-watermark-audit.git
```

For development:

```bash
git clone https://github.com/steven-panxd/ai-text-watermark-audit.git
cd ai-text-watermark-audit
python -m pip install -e .
```

## Use

Scan a UTF-8 file:

```bash
textmark scan draft.txt
```

Produce JSON and fail CI on warning-level findings:

```bash
textmark scan draft.txt --json --fail-on warning
```

Compare two versions:

```bash
textmark compare original.txt edited.txt --json
```

Show sourced vendor claims and detector availability:

```bash
textmark claims
```

The scanner currently reports:

- zero-width and format characters;
- bidirectional text controls;
- Unicode tag characters;
- unusual spaces and variation selectors;
- mixed Latin/Cyrillic or Latin/Greek tokens;
- exact positions, Unicode names, escaped values, and SHA-256 hashes.

It runs locally, has no runtime dependencies, and does not upload text.

## What a finding means

A hidden character is not proof that text was generated by AI. Many invisible characters have legitimate uses in typography, emoji, and non-Latin writing systems. Conversely, a clean scan does not prove human authorship and does not rule out a statistical watermark.

The report is forensic input, not an authorship verdict.

## Current Claude status

As of 14 August 2026:

- Anthropic says Claude models launched on or after 2 August 2026 support machine-readable marking at launch.
- It says older models are being updated.
- Marking applies worldwide on supported models.
- Anthropic has not yet published technical documentation for third-party detection.

Primary source: [How Claude marks AI-generated content](https://support.claude.com/en/articles/16266773-how-claude-marks-ai-generated-content).

## Roadmap

- A stable detector-adapter interface for official vendor tools
- Corpus manifests and reproducible robustness runs
- File provenance inspection for supported formats
- Exportable Markdown and SARIF reports
- Multilingual false-positive fixtures

## Development

```bash
python -m unittest discover -s tests -v
```

See [CONTRIBUTING.md](CONTRIBUTING.md) before proposing a detector or vendor claim. Security-sensitive findings should follow [SECURITY.md](SECURITY.md).

## License

[MIT](LICENSE). This project is independent and is not affiliated with Anthropic or other model providers.
