A video is not an image.
That is why searching one is hard.
Every reverse image search wants a still. A film is thousands of them.
There is no single button for this, and any page that says otherwise is selling something. What follows is how reverse video search actually works, which tool to reach for depending on what you are trying to find out, and where STEALSHOT fits — 1,038 launch and demo films, already cut into 90,425 frames and measured.
- The frame method, step by step, with the part everyone gets wrong.
- Five tools compared by what each one can actually find.
- Why Google has no reverse video search, and what it has instead.
- A library of launch films you can search without extracting anything.
Free, no account needed to search
How to reverse search a video
Every general-purpose method comes down to the same move: stop treating it as a video. Reverse image engines index stills, so the job is to pick the right still and search that.
- 1Pick the frame
Not the thumbnail, and not the first frame — both are usually a title card that appears in a thousand other videos. Scrub to somewhere distinctive: a face, a location, an object, on-screen text. This choice decides the whole result and it is the step people skip.
- 2Capture it clean
Pause and screenshot, then crop away the player chrome, the progress bar and any watermark the platform burned in. A reverse image engine matches on the whole picture, so a YouTube play button in the middle of your crop is noise it has to search through.
- 3Search the still
Run the crop through Google Lens, then repeat it in Yandex and TinEye. They index different corners of the web and disagree constantly, which is the point — three searches on one good frame beats one search on three bad ones.
- 4Try more frames
A miss means the frame was wrong, not that the video is unfindable. Pull three or four stills from different scenes and run each. If a video has been reuploaded anywhere, one of them usually lands.
Which tool finds what
None of these searches a video. They search a picture you took of one — which is why what they find differs so much.
Objects, places, products and text inside the frame. The widest index of the five, and the one to try first.
Matches what is in the picture rather than the picture itself, so it often returns things that merely look similar.
Faces and exact-crop matches. Routinely finds reuploads the others miss, and is the strongest of the five on people.
Coverage skews to the Russian-language web, and the face matching raises obvious questions about who you point it at.
Provenance. Sorts by oldest first, which is the only way to answer who posted it before everybody else did.
Exact and near-exact copies only. A re-encode, a crop or a filter can be enough to lose it.
Doing the frame extraction for you. Built for newsroom verification: paste a link, get keyframes, search each one.
A browser extension aimed at journalists, not a search engine. It prepares the search; it does not answer it.
Launch films, product demos and announcements. Already indexed frame by frame — search it by URL, brand or what is happening on screen.
One category, deliberately. It knows 1,038 films inside out and nothing about the rest of the web.
What STEALSHOT does, and what it does not
It is worth being plain about this, because most of the traffic to a page like this one is looking for something else. If you are trying to identify a stranger, trace a rumour or check whether a clip is real, the tools above are the right ones and STEALSHOT is not.
What it does is narrower and much deeper. Every launch and demo film that goes in gets cut into frames, and every frame gets described and measured — where the presenter sits, how fast it cuts, how dark it is, what the type does. So the question it answers is not where did this come from. It is: here is a video, what came out of it — and what else looks like this.
Which makes it useful if you are about to shoot one of these and want to see how it has been done 1,038 times already.
Questions
Does Google have a reverse video search?
No, not as a single feature. Google Lens searches a still image, so the standard workaround is to screenshot a frame and search that. Google has never shipped a search that takes a video file and matches it against video on the web.
How do I reverse search a video on my phone?
The same way, with fewer steps. Screenshot the frame, open Google Photos or the Lens app, and search the shot from your camera roll. On iPhone, long-pressing the subject in a screenshot gives you Visual Look Up, which is Lens by another name. Cropping the player controls out still matters.
Is reverse video search free?
Google Lens, Yandex, TinEye and InVID are all free. Paid services mostly resell the same engines with monitoring on top. Searching STEALSHOT is free and needs no account; pulling a film down is three a month free, uncapped on Pro.
Can I reverse search a TikTok or Instagram video?
Yes, with the frame method — screenshot, crop, search. What no tool does well is search inside those platforms, because neither lets reverse image engines index their video. Expect to find reuploads on the open web rather than the original post.
Reverse video search or reverse image search — what is the difference?
In practice, one step. Reverse image search takes a still and asks where else it appears. Reverse video search is the same engine with a frame-extraction step bolted on the front. That is why results vary so much with which frame you pick.
How do I find the original source rather than a reupload?
Sort by date. TinEye's oldest-first sort is built for exactly this and is the fastest route to a first appearance. Lens and Yandex rank by similarity, not by time, so the top result there is often a popular repost rather than the origin.
Or skip the frame extraction entirely
1,038 launch films, 90,425 frames, already pulled apart and measured. Search by brand, by what is on screen, or by pasting a link.