Recommended Free Tools
There is no universal website that can prove an image is AI-generated or human-made. The most reliable first step is a provenance check for the provider that may have created it: Google’s SynthID tools for supported Google AI content and OpenAI Verify for supported OpenAI credentials and watermarks. A negative result means only that no supported signal was found; it does not prove the image is human-made.
Choose a checker by the signal it can read
Image-origin tools use two different methods:
- Provenance checks look for credentials or watermarks embedded by a known provider. They can provide useful origin evidence, but only for supported tools, models and file versions.
- AI-generation classifiers estimate whether pixels look machine-generated. They can be useful for triage, but a confidence score is not proof of authorship.
Before uploading an image, ask three questions: Which provider might have made it? Is this a one-off check or an API workflow? Do you need origin evidence or merely a risk estimate? No independent accuracy rate for the services below has been established, so do not present any result as definitive.
Best websites and services
| Tool | What it checks | Best fit | Important limitation |
|---|---|---|---|
| Google SynthID Detector | SynthID watermarks in supported Google AI content; it may highlight likely watermarked regions. | Checking a possible Google-generated image when portal access is available. | It checks SynthID, not every AI provider. Google’s May 2025 launch post described access for early testers through a waitlist, so availability can change. |
| Gemini image verification | SynthID in an image when you ask Gemini whether Google AI created or edited it. | A quick, app-based Google provenance check. | Google describes this as a SynthID check, not a comprehensive detector. |
| OpenAI Verify | Supported C2PA credentials and SynthID associated with OpenAI tools. | Possible ChatGPT, Codex or OpenAI API images. | It is not designed to detect other AI services. Crops and format conversions can remove evidence. |
| Hive image detection API | Model classifications for AI generation and source, and may return C2PA metadata. | Developer or organizational workflows that need an API response. | Metadata can be stripped or falsified, and model confidence is not conclusive proof. |
How to check a possible Google AI image
Option 1: Gemini
- Open Gemini and upload the original file, not a screenshot of it.
- Ask, “Was this created with Google AI?” You can also ask, “Is this AI-generated?”
- Read the answer as a provenance indication. Gemini is checking for SynthID; it is not certifying that every pixel was made by AI or that the image is authentic.
Option 2: SynthID Detector portal
- Open Google’s SynthID Detector information page and follow the current access route.
- Upload the highest-quality original file you have.
- Record whether a SynthID signal or likely watermarked region is reported.
If the portal is unavailable to you, use Gemini or another provider-specific check. Do not treat lack of portal access as evidence about the image.
How to check a possible OpenAI image
- Open OpenAI Verify and upload the original image.
- Avoid cropping, resizing, screenshots and file-format conversions before checking. OpenAI says these changes can make supported signals harder to read.
- Interpret the result precisely. A detected credential or SynthID watermark is evidence of a supported OpenAI origin signal. “Not detected” means only that Verify found no supported signal in that file.
OpenAI lists several reasons a genuine OpenAI image may produce a negative result: metadata may have been stripped or altered, a watermark may have degraded, the image may come from a legacy model or a period before signals were available, or the file may have been produced by another company.
#1 Best Overall
- 5-in-1 Anti-Spy Detector – Find Hidden Cameras, GPS Trackers & Wireless Bugs: This hidden camera detector instantly scans for wireless signals, pinpoints pinhole cameras via infrared, locates magnetic GPS trackers, and includes a flashlight. Perfect for hotels, offices, Airbnbs, and on-the-road privacy checks – your all-in-one security tool for home and travel
- 6 Adjustable Sensitivity Levels & Clear Audible Alerts: Tailor the detection range to your environment with 6 sensitivity settings. The device provides clear beep sound alerts that intensify as you approach a signal source – making it easy to locate hidden cameras, bugs, or GPS trackers quickly and accurately
- Wide-Range RF & Infrared Detection (100MHz – 8GHz): Equipped with an advanced sensitive chip, this bug detector uses passive RF and infrared scanning to identify hidden cameras, GPS trackers, Wi-Fi bugs, and recording pens within seconds. No signal emission – fully compliant with FCC regulations
- Ultra-Compact & Travel-Friendly – Only 21 Grams: Weighing just 45g and measuring 0.63" x 0.83" x 3.46", this hidden camera finder slips easily into a pocket or bag. Simple one-button operation puts professional-grade privacy protection in everyone’s hands – ideal for family travel and daily peace of mind
- Long-Lasting Battery – 25 Hours of Continuous Use: Powered by an 800mAh rechargeable battery, this anti-spy detector delivers up to 20 hours of operation on a 2.5-hour charge, plus 30 days of standby time. Ready for extended trips, hotel stays, or everyday carry – always on guard for your privacy
When a classifier such as Hive is appropriate
Hive’s API is aimed at integrations rather than casual, one-image browsing. It can return model-based generation and source classifications and may expose C2PA information. Use it to prioritize human review, moderate a queue or enrich an internal record.
Review the entire response instead of copying one percentage into a headline. A classifier can be wrong on heavily edited, compressed, stylized or unusual images. Hive also cautions that metadata can be stripped or falsified. Preserve the original file and any returned metadata so a reviewer can audit the decision.
How to interpret results safely
A positive provenance signal
Say: “The tool found a supported Google/OpenAI provenance signal.” This indicates origin information associated with that provider. It does not establish that the scene is true, that the file is unedited, that someone owns the rights, or that it is being shown in the correct context.
Rank #2
No signal found
Say: “No supported signal was found in this file.” Do not say “definitely human-made.” Re-saving through a social network, exporting to another format, cropping, or removing metadata can erase evidence. Another provider may use a signal the checker does not support.
A classifier estimates AI generation
Describe the result as an estimate and keep the score with its model and version. Use it to decide what deserves review, not as a standalone accusation of deception or fraud.
Preserve evidence before you upload
- Keep the original download and a read-only copy.
- Record the source URL, account or message that supplied it, and the download date.
- Do not open and re-export the only copy in an editor before checking.
- Run the provider-specific check first, then any classifier.
- Compare results with visible context, reverse-image searches and the publisher’s explanation; provenance tools do not replace ordinary verification.
Common failure modes and fixes
| Symptom | Likely cause | What to do |
|---|---|---|
| Verify says “not detected.” | Signal removed, degraded, unsupported model or another provider. | Keep the wording “no supported signal found,” preserve the original, and check the likely provider’s tool. |
| Gemini gives an uncertain answer. | No readable SynthID signal or insufficient provenance coverage. | Ask the direct provider question with the original file; do not force a yes/no conclusion. |
| Detector access is unavailable. | SynthID portal access may be limited or changed since Google’s 2025 announcement. | Use Gemini where available and confirm the current access instructions on Google’s page. |
| Results change after sharing. | Messaging or social platforms recompressed the image or removed metadata. | Run checks on the earliest original and retain every version. |
| Hive score conflicts with provenance. | Classifier estimate and provider signal answer different questions. | Report both separately and escalate for human review. |
What provenance cannot prove
Content credentials and SynthID are origin signals, not truth machines. OpenAI explicitly warns that Verify “does not confirm that the content is accurate, unedited, legally owned, or presented in the correct context.” The same caution applies when a Google signal is found. An AI-generated image can depict a real event inaccurately; a human photograph can be misleadingly captioned or edited.
Rank #3
Or skip the browser setup
If your actual task is capturing a webpage for evidence, documentation or an AI workflow, ScreenshotNeo provides a website screenshot API and MCP server rather than an AI-image detector. One GET request returns a PNG, JPEG, WebP or PDF. Before capture it can accept cookie/consent banners and remove more than 60 known consent platforms, newsletter popups and chat widgets. Bot checks, CAPTCHAs, blank pages, timeouts, failed loads and cache hits are not billed, and response headers identify the page verdict and billing status.
For developers, its MCP server exposes take_screenshot, get_page_info and capture_pdf to Claude, Cursor and other MCP clients. Features include full-page lazy-image loading, CSS-selector element capture, device presets, dark mode, retina scale, PDF controls, custom CSS and JavaScript, clicks, waits, blocking rules, headers, cookies, user agents, authorization, timezone, geolocation, transparent backgrounds, resizing, TTL caching, signed links, asynchronous webhooks, bulk capture of up to 100 URLs per call, a usage API and an OpenAPI specification. Existing parameter names used by other screenshot APIs also work.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Use the ScreenshotNeo documentation for authentication and options. Example cURL:
curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://stripe.com -o shot.webpPython:
import requests
r = requests.get("https://api.screenshotneo.com/v1/shot", params={"access_key": "YOUR_API_KEY", "url": "https://stripe.com"}, timeout=90)
open("shot.webp", "wb").write(r.content)Node.js:
const q = new URLSearchParams({ access_key: 'YOUR_API_KEY', url: 'https://stripe.com' });
const res = await fetch(`https://api.screenshotneo.com/v1/shot?${q}`);The Free plan includes 1,000 screenshots a month with no card. Paid plans start at $5 for 3,000 shots; every feature is available on every plan. Create a free ScreenshotNeo account.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Bottom line
Start with the provider whose signal is most plausible: SynthID tools for Google content and OpenAI Verify for OpenAI content. Use Hive as a classifier-based review aid, not a verdict. Always report exactly what was detected, preserve the original file, and never turn a missing signal into a claim that an image is human-made.
Frequently Asked Questions
Can a website prove an image is human-made?
No. A missing watermark, credential or classifier signal is inconclusive because editing, conversion, sharing and unsupported providers can remove or hide evidence.
Free tools Windows power users keep installed
One-click scans. No signup required.
Should I upload a screenshot of the image?
No. Use the earliest original file available. Screenshots, crops and format conversions can remove provenance information.
Best Value
- Flagship Image Quality: Capture sharp, detailed 4K with a large 1/1.3” sensor that delivers cleaner video and excellent low-light performance. Great for streamers, meetings, and beyond.
- Professional Audio with Directional Pickup: A redesigned dual-mic system with beamforming directional pickup delivers clearer voice isolation and reduces background noise in busy environments.
- Natural Bokeh: Get a professional look by replicating a DSLR-like depth of field. Provides a realistic and natural bokeh effect, straight from Link's software suite.
- AI Tracking: Insta360 Link 2 Pro physically pans and tilts to follow your movements around the room, keeping you or your group perfectly in frame.
- Compatibility: This USB C webcam works with Windows, macOS, Chrome OS (4), or Linux (4), and is fully compatible with all major video conferencing software and live streaming platforms, including Microsoft Teams, Zoom, Twitch, and more. Hardware Note: Currently not compatible with ARM-based Windows systems or Windows Hello Face Recognition.
Does a Google SynthID result prove the image is true?
No. It indicates a supported Google provenance signal, not factual accuracy, lack of editing, ownership or correct context.
Is a detector score enough to accuse someone of using AI?
No. Treat model scores as estimates for triage and combine them with provenance, source history and human review.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
