Tool suggestion: isGenAI — saved AI image details and prompts

Affiliation: this suggestion comes from the isGenAI project and was prepared with coding-assistant help.

Tool: [isGenAI image checker](https://isgenai.com/ai-image-detector)

Suggested category: Image/Video → Metadata.

For a researcher checking an original image, isGenAI reads saved AI generation details, supported AUTOMATIC1111/ComfyUI prompts and available signed creation records. The result identifies the details found and checks run, and can be saved as a text report. The saved-prompt workflow covers supported PNG records and AUTOMATIC1111 settings in JPEG/WebP; Flux/ComfyUI workflows with multiple stages can be identified without claiming a single recoverable prompt.

Access: free web workflow, no account required, up to 20 MB per image. Server upload is required. The [privacy policy](https://isgenai.com/privacy) explains processing and retention. Public-URL checks can create public receipts; uploaded files are private by default. Requesting Get prompt can send content to DeepSeek for a new suggested prompt if no supported saved prompt exists. Researchers should review those processing choices before submitting sensitive material.

Limits: saved metadata can be edited or stripped. The service does not establish whether an event or caption is true, and missing metadata does not establish human origin. We are suggesting it for inspecting recorded details, not as a generally accurate human-versus-AI classifier.

Reproducible example: the [image guide](https://isgenai.com/guides/how-to-check-ai-image#examples) has two downloadable copies of the same picture, with and without deliberately added saved details, plus method notes and result files. The example is explicitly labeled as a metadata demonstration rather than an accuracy benchmark. The [API guide](https://isgenai.com/guides/ai-detector-api) provides a reusable Node.js check script.

No existing isGenAI entry or earlier issue was found. This is a tool suggestion for editorial assessment; it does not imply toolkit inclusion or endorsement.

Start by reviewing the linked image guide, API guide, privacy policy, and reproducible examples. Check whether the tool fits the toolkit's editorial standards and proposed Image/Video → Metadata category, including its limitations and processing disclosures. Done means an editorial inclusion or rejection decision with the relevant entry requirements identified.

Written by the indexing model from the issue text.