Raid AI is now on X: tag @Raidxai under any post and get the verdict

A convincing fake does not stay where it was posted. It is quoted, reposted and argued about, and by the time anyone checks it, thousands of people have already made up their minds. Raid AI is now on X, so the check can happen where the post is: mention @Raidxai under it and the verdict comes back in the same thread.

1. 01Find the post with the video, image or voice you are not sure about.
2. 02Reply to it with @Raidxai and a question, such as "@Raidxai is this real".
3. 03Raid AI replies in the thread with its verdict, and X notifies you when it does.

You do not have to reply to the post itself. If you mention @Raidxai further down a conversation, it walks back up the thread to the first post that carries media and checks that one. When the media sits more than one post above yours, the verdict goes on that post and tags you, so everyone reading the original sees it too.

The verdict comes first, in plain words, so it reads at a glance in a busy timeline:

• —Likely AI generated 🔴 when the evidence points to a generated or manipulated post.
• —Likely real 🟢 when nothing in it looks generated.
• —Not sure 🟡 when the evidence is split, rather than a guess dressed up as certainty.

Under it comes one line for each part of the post that was checked, video, image or audio, with the flagged part first, for example signs of AI in the video while the audio looks real. There are no percentages in the reply. A confidence score is easy to misread in a timeline, and the verdict is what people act on.

Each format is judged by several independent detectors, including Raid AI's own models, and each format gets its own verdict, so a real voice laid over a generated video is still caught. It is the same detection that runs in the Raid AI platform for banks, insurers and call centres, answering in a place where anyone can use it.