Get a simple result first. Open the details to see exactly what we found, where it came from and how we checked it.
About
About isGenAI, the free AI detector at isgenai.com
isGenAI (isgenai.com) is a free AI detector for images, text and code that shows the evidence behind each result: C2PA, AI labels and saved prompts. It reads AI labels, saved prompts, creation records and matching copies, and every report separates what the content proves from what it leaves open. Start instantly, no account needed.
At a glance
The product is an independent web application and API that checks content for evidence of AI use.
- Name
- isGenAI, at the canonical domain isgenai.com
- What it is
- A free AI detector and evidence checker for images, text, code and public links, with free browser-based image tools and an HTTP API.
- Founded
- 2026; the public checker went live at isgenai.com on .
- Who runs it
- An independent team that builds and operates the checker and writes its guides. The product is not affiliated with OpenAI, Anthropic, Google, Adobe or the C2PA.
- What it reads
- C2PA Content Credentials, the IPTC digital source type, saved prompts from AUTOMATIC1111, ComfyUI and InvokeAI, generator metadata and AI-use statements in text.
- What it does not do
- It gives no visual-only verdict, no writing-style authorship score and no accuracy percentage; a missing AI signal is never presented as proof of human authorship.
- Topics
- AI image detection, AI-generated content, C2PA Content Credentials, IPTC digital source type, AI image prompt metadata (AUTOMATIC1111, ComfyUI, InvokeAI), AI provenance, AI watermarking, AI-use disclosure in text, EU AI Act Article 50, synthetic media disclosure.
- Contact
- Contact page · Corrections · Independence policy
How a check works
A check starts from the file, text or link you supply. For images, the reader opens the file’s own records: a C2PA manifest and its signature, the IPTC digital source type, prompts and settings saved by AUTOMATIC1111, ComfyUI or InvokeAI, and other generator metadata. For text and code, it looks for sentences that declare AI use and for named tools such as ChatGPT, Claude, Gemini, DeepSeek and Copilot. Each finding in the report names the field or sentence it came from, and the report lists which checks completed. Screenshots, re-exports and social media uploads often remove records, so a clean result describes that copy only. The report never turns the look of a picture or the style of a sentence into a verdict.
The isGenAI checker is published only at isgenai.com. Similarly spelled names on other domains are separate and not affiliated with this product.
Images carry saved details about the tools and edits that made them. We read those details, check AI labels and connect matching copies in our records, all from the file itself.
Explore our Image Metadata Lab: download controlled files, verify their fingerprints and repeat every check. Read the developer guide to run the same readers in your own workflow.
We test checks against research datasets with known origins before anything ships. Your feedback decides which problems we investigate and test next.
Email h.a@skyvaults.io with questions. Our correction process shows how to submit context and verifiable evidence.
Run the same checker through our Apify Actor, RapidAPI listing or find us on Product Hunt. Every listing points to the same checker at isgenai.com.
Who publishes these tools and guides?
isGenAI publishes the tools, worked guides and controlled research at isgenai.com. An isGenAI byline is our own product documentation and experiments, written by the people who build the checker. It is not an independent review, a professional accreditation or an endorsement from a standards body, and we say so plainly.
Check the work behind every answer
Our worked guides walk through one specific task each and link to the relevant tool. The Image Metadata Lab supplies original files, recorded outputs, hashes and reproduction code. The signed and unsigned C2PA examples show exactly why a signature result and trust in its signer are read separately.
These controlled examples demonstrate real file behavior, reproducibly. They are file-behavior studies, distinct from accuracy benchmarks or proof that a real photograph is genuine. Read the checker methodology and limits before using a result to make a decision about someone else’s work.
Sources, updates and corrections
Technical guides link to source documentation and carry their publication or review date, which we update whenever the content changes substantively. External platforms change, so repeat any example with your actual files to confirm it still holds.
Found a wrong explanation, an unreadable file or an unexpected result? Tell us what you tried, or use the correction process for an existing report. Give the page or report link and the smallest reproducible example you can share.
Last updated by isGenAI.