Article
Images can carry two very different kinds of ideas. One might be a forwarded screenshot with a claim in it. Another might be a picture you want to create from a written description. VOM handles both: it can read and check the text in an uploaded image, or turn your idea into a generated image.
Check a forwarded screenshot without retyping
When a claim arrives as a screenshot, the first step is getting the words out of the image. Send the screenshot to VOM, and it reads the text inside it. You do not need to retype the message before checking it.
VOM then returns the text it extracted from the image. This gives you a chance to review what it read and edit the text before anything is checked. If the screenshot includes wording that was read incorrectly, or if you want to adjust the text being checked, you can do that at this stage.
That review step keeps the image and the words visible as separate parts of the process: the screenshot is the input, and the extracted text is what you can inspect and edit before verification.
See what the image check returns
After the text has been reviewed, VOM assesses the uploaded image. Its assessment places the image in one of three categories: likely authentic, likely AI-generated, or unverifiable.
The result includes a labeled verdict, a confidence score, and the sources behind it. The label tells you which assessment VOM returned. The confidence score appears with that verdict, while the visible sources show what sits behind the result.
This gives the check a visible result to examine rather than leaving the uploaded screenshot as an unreviewed claim. You can see the extracted wording, make any needed edits before the check, and then view the verdict, score, and sources returned for the image.
Turn a written idea into an image
VOM also works in the other direction. Instead of starting with an image you received, start with a written idea. Describe what you want, and VOM turns that idea into a generated image.
The starting point is your description. The output is an image generated from that written idea, giving you a visual to work with without beginning from an existing picture.
Once the image is generated, signed-in users can save it. They can then download or share it. That makes the generated image available beyond the generation screen, whether you want to keep a copy, use the downloaded file, or send the image to someone else.
Two image tasks, one workflow
The checking side begins with an image and moves toward text: upload a screenshot, review the extracted wording, edit it if needed, and see the image assessment with its labeled verdict, confidence score, and sources.
The making side begins with text and moves toward an image: describe an idea, receive a generated image, and, when signed in, save it before downloading or sharing it.
The two paths answer different image needs. A forwarded screenshot can be brought into a check without retyping its text. A written idea can become a generated image that you can save, download, or share when signed in.
Before a screenshot is checked, the extracted text is there for your review. Before a generated image leaves VOM, signed-in users have the option to save it and then download or share it.