Independent SynthID guide & reports
Learn how Google’s SynthID watermark works, how to verify media with the official free detector, and when a structured report helps journalists, marketers, and creators move faster.
Free checks run on Google’s official tool (~10/day with login). We explain results and offer paid reports — we do not run Google’s detector ourselves.
SynthID is an imperceptible watermark technology from Google DeepMind. When supported AI systems generate or edit media, SynthID can embed a signal into images, video, audio — and in some cases text — that a detector can later look for.
It is not a universal “is this AI?” detector. A “not detected” result does not prove human authorship. Content from unsupported models, heavily re-encoded files, or systems that never apply SynthID may show no watermark even when AI was involved.
Google has opened SynthID Detector as a global portal. Coverage includes media watermarked by participating partners (e.g. Google, OpenAI, NVIDIA, Kakao; Apple announced as coming). Always treat the official detector’s current documentation as the source of truth for formats and partners.
We help you understand SynthID, use the official free checker correctly, and — if you need more than a few spot checks — get a batch report, journalist checklist, or clear result interpretation. Start with our What is SynthID guide or jump to all guides.
The free detector lives at synthid.com. Use it for individual checks. Follow these steps:
Open synthid.com and sign in with Google, OpenAI, or Apple (as offered on the official site).
Submit the image, video, or audio file you want to verify. Prefer the original or least-compressed version you have.
Note whether SynthID was detected. Save a screenshot or export notes for your editorial or brand workflow — especially if you need an audit trail later.
Need to verify many files, document findings for a newsroom, or interpret ambiguous outcomes? See our Batch Report — built for that workflow.
| Result | Practical meaning |
|---|---|
| Detected | The file appears to carry a SynthID watermark from a supported AI system (generated or edited). The official tool does not always separate “fully generated” vs “edited.” |
| Not detected | No SynthID watermark found. This does not prove the media is human-made. It may be from an unsupported model, stripped/altered, or never watermarked. |
| Unclear / inconclusive | Treat as “needs more context.” Check file provenance, ask the source, and avoid overclaiming in publication. |
Google’s free detector is useful for spot checks, but accounts are typically limited to about 10 checks per day (login required). That is enough for a few stories or campaign assets — and a bottleneck for:
Our paid options exist for that gap: structured reports, checklists, and interpretation — not a fake local detector pretending to be Google.
Detection always runs on Google’s official tool. Pay only when you need a filled report or editor checklist — preview samples before you order.
Payments coming soon — email to order. Stripe is not live yet. Use a product button to email upgrow.us@gmail.com with the product name and your files or links.
Free
$0 on synthid.com
Google’s free watermark checker for individual files.
What you get
Most popular
$9.99 one-time
A filled verification report for your file list — emailed after checks on the official Detector.
What you get
Sample deliverable
Batch Report
See full sample →Lightweight
$4.99 one-time
PDF checklist + result templates — process and language, not a filled multi-file report.
What you get
Sample deliverable
Editor Quick Guide
See full sample →Note: Free detection is on Google’s SynthID Detector. Paid products here are workflow documents (report / checklist / interpretation) — not Google’s engine, and not a public Detector API.
Short, practical articles for editors and marketers. Free detection still happens on the official tool at synthid.com.
How Google DeepMind’s imperceptible AI watermark works — and what it does not prove.
Read guide →Step-by-step: sign in, upload media, read results, and work within the daily limit.
Read guide →Why watermark checks differ from “is this AI?” classifiers — and when to use each.
Read guide →See the exact deliverable format: file table, Detected / Not detected rows, and interpretation notes.
Preview sample →Editor checklist and copy templates included in the $4.99 PDF — preview before you buy.
Preview sample →No. checksynthid.com is an independent guide and paid-report site. The free detection tool is operated by Google at synthid.com. We are not affiliated with, endorsed by, or part of Google.
The official tool is the right place for individual free checks (typically ~10/day after login). Pay only if you want a batch report, a journalist checklist, result-interpretation help, or a documented workflow beyond quick spot checks. Preview deliverables: sample Batch Report and sample Quick Guide.
Documents — not a private detector. Batch Report = filled file table + notes + language templates emailed to you. Quick Guide = PDF checklist and interpretation templates. Detection still happens on Google’s official SynthID Detector. See the samples linked under Pricing.
The official detector supports images, video, and audio watermarked with SynthID by participating systems. Exact formats and partner coverage can change — confirm on synthid.com before you rely on a check for publication.
No. Absence of SynthID does not prove human authorship. Many AI tools never apply SynthID; files can also be altered or recompressed. Use watermark checks as one signal among provenance, source interviews, and editorial judgment.
We do not provide Google’s SynthID Detector API. Google’s consumer detector is web-based with login and daily quotas; a general public Detector API for consumers was not available at the time of writing. Our paid offerings are report/checklist products, not API access to Google’s engine.
For paid report requests, we only use files to complete the verification workflow you ordered, then follow our retention practice (see Privacy). Do not upload sensitive material you are not authorized to share. Prefer links or samples when possible if your policy requires it.
Journalists and editors verifying media before publish; marketers and brand teams reviewing campaign assets; creators and agencies who need a clear process for AI watermark checks — without hype or fake confidence scores.