# AIGCTrace

> AIGCTrace is a free AI-generated image provenance and watermark detection tool for checking whether an image contains signals of AI generation, C2PA Content Credentials, or visible/invisible watermark evidence.
>
> AIGCTrace 是一款免费的 AIGC 图片溯源与水印检测工具，帮助用户判断图片是否含有 AI 生成来源线索、C2PA 内容凭证或可见/不可见水印证据。

## Site Languages

- Primary URL: https://aigctrace.com/
- Supported languages: English and Simplified Chinese
- Language behavior: the same web app supports both languages through an in-page language switcher and browser-language detection.
- 站点语言：英文、简体中文
- 语言机制：同一个 Web 应用通过页面内语言切换和浏览器语言识别提供双语界面。

## Product Summary

AIGCTrace analyzes images for AI-generation provenance signals across C2PA metadata, visible watermarks, invisible digital watermark clues, blind watermark patterns, and common AIGC metadata fields. Best support for ChatGPT, Gemini, Doubao, and Qwen; covers 46 AI tools. Supports TC260 / GB/T 45438-2025 China AIGC identification standard. Returns a risk score, risk level, confidence values, key evidence, and a downloadable share card.

AIGCTrace 通过多维检测技术分析图片中的 AI 生成痕迹，覆盖 C2PA 内容凭证、明显文字水印、隐形数字水印线索、盲水印结构和常见 AIGC 元数据字段。重点支持 ChatGPT、Gemini、豆包、通义千问等主流工具，累计覆盖 46 种 AI 工具。支持 TC260 / GB/T 45438-2025 中国 AIGC 标识标准。输出综合风险评分、风险等级、置信度、关键证据和可下载分享图。

## Core Features

- **C2PA Content Credentials detection / C2PA 内容凭证检测**: detects C2PA-compatible provenance metadata, AIGC tool markers, XMP/IPTC/EXIF fields, PNG text chunks, and AIGC-related JSON markers. Best support for ChatGPT, Gemini, Doubao, and Qwen; covers 46 AI tools. Also supports TC260 / GB/T 45438-2025 China AIGC identification standard.
- **Invisible digital watermark detection / 隐形数字水印检测**: checks for public AI-origin clues and invisible watermark-like signals embedded in image data, including SynthID, filename markers, and IPTC tags.
- **Blind watermark clue detection / 盲水印线索检测**: analyzes low-bit-plane periodic structures and repeated texture-like signals that may indicate blind watermarking.
- **Visible text watermark detection / 明显文字水印检测**: estimates whether edge regions contain high-contrast text-like overlays, visible marks, or platform labels.
- **Risk report / 风险报告**: provides a 0-100 score, low/suspicious/high-risk labels, per-detector status, confidence percentages, evidence details, and capability limits.

## Supported Formats

- PNG
- JPEG
- WebP
- GIF
- TIFF
- Maximum file size: 15 MB
- 最大文件大小：15 MB

## Privacy

- Original images are not stored.
- Plaintext filenames are not stored.
- IP addresses and User-Agent values are stored as salted hashes by default.
- Only coarse context such as country, region, city, ASN, and referer may be logged for usage and abuse analysis.
- 不保存原图。
- 不保存明文文件名。
- IP 和 User-Agent 默认以 salted hash 存储。
- 仅记录国家、地区、城市、ASN、referer 等粗粒度上下文，用于使用统计和滥用分析。

## Use Cases

- Media and content teams verifying image provenance and authenticity.
- Creators checking whether their work carries AI-generation labels or watermarks.
- Platforms and moderation systems using AIGC signals as auxiliary review evidence.
- Security researchers evaluating AIGC watermarking and provenance technologies.
- 媒体和内容团队核查图片来源真实性。
- 创作者验证作品是否带有 AIGC 标记或水印。
- 平台和审核系统辅助判断 AI 生成内容。
- 安全研究人员分析 AIGC 水印与溯源技术。

## Key Resources

- [Homepage and landing page](https://aigctrace.com/): Canonical landing page for the free bilingual AIGC image provenance detection web app.
- [llms-full.txt](https://aigctrace.com/llms-full.txt): Detailed bilingual technical and product context for AI agents.
- [Sitemap](https://aigctrace.com/sitemap.xml): Canonical sitemap.
- [Robots policy](https://aigctrace.com/robots.txt): Crawling policy for search engines and AI crawlers.