Vidit reads the geolocations you already publish on X and creates a detection for each one, with its coordinates, source, media and proof note. There are three ways in and one engine behind them, so a post reads the same whichever you use. There is no format to learn: write your post as you always do.
Two conditions, the same for every entry. The post is yours: it comes from the X handle linked to your Vidit account, and a retweet does not count. It carries a coordinate: anywhere in its text, or in the text of your own post it directly replies to. A coordinate that appears only in a post you quote does not count.
A detection is on the map from the moment it lands, marked as a machine detection and attributed to your account, and it waits in your detections queue. Review it, correct the event date, then publish it as a geolocation. Rejecting it removes it. Only you turn your detection into a geolocation.
A post carrying several coordinates makes one detection each, sharing every other field. Importing the same geolocation again reuses the first detection instead of duplicating it, and a detection you already published or rejected stays as it is, so re-running an import is always safe.
Upload your official X export and every geolocation you already published comes back as a detection. This is the only entry that reads your whole history at once, and the only entry that stitches full self threads, so a coordinate you posted three replies down still lands beside the source you posted at the top.
Request your archive on X
Wait for X to build it
Download the .zip
Upload it on Vidit
The stripped zip must stay under 4 GB. Contact us if yours is larger after stripping. X's own guide
An export runs as a background job, so you can close the page. Vidit emails you the outcome when it finishes: how many detections were created, updated and skipped, and how many need a source or carry several coordinates. Uploading the same export again creates no duplicates, which is also how you resume an import that failed.
Open Submit, pick From an X post and paste the link to one of your own posts. The detection is created while you wait; the review opens on it when nothing needs your attention, otherwise the page tells you what does. Your own posts only, matched against the X account linked to your profile; a third party's footage goes through the plain submit form with a source URL.
Warnings and refusals are shown on the page rather than on X, and nothing is posted under your handle, so this is the way in when you do not want a public trace. Pasting a post the bot already imported reopens that detection instead of making a second one.
Tag @ViditBot on a geolocation post and the detection appears on Vidit within seconds. You do not leave your feed and you retype nothing. The bot imports only for X handles linked to a Vidit account: it stays silent for any other handle and creates nothing. It reads public posts only, so a tag from a protected account imports nothing.
Post the geolocation, then reply to yourself with the source and the tag. The bot reads both posts as one, and the detection is filed under the first of them, so tagging either one imports the same geolocation once. This is also how you import a post you published earlier. The coordinate, the source link and the tag can also sit on one post.
analyst
@analyst
analyst
@analyst · replying to @analyst
The bot answers in-thread with your detection's reference and with what to fix at review: an empty source, several coordinates, a missing footage file or post date, or media already on Vidit. When nothing imports, it names which of the three refusals it was. Replies are capped per hour, so a burst of tags still creates every detection while some of them go unanswered.
Vidit
@viditbot · replying to @analyst