Bots, fakes and scammers, gone.
Bot Purge scans your followers, friends, DMs, texts, email, connected apps and live chats, shows you every threat it finds, and removes them. For people, creators and whole platforms.
Clean up your accounts
Free scan. You never need to give us your social media passwords: connect X with its official sign-in, or upload the data export each platform lets you download.
Scanning a teen's account as a parent or guardian?
Only with the teen's knowledge and consent. Enter their handle and confirm below.
Already have an access key?
Purge bots from your whole platform
Score every account on your service, remove bot rings in reversible tiers, and give every affected user a notice and a fair appeal. Full audit log, dry runs, rollback, a real-time signup gate, and Discord and Discourse integrations.
Open the Purge ConsoleGet the app
The desktop app keeps everything on your computer, reads your iPhone text backups, and runs Live Guard and the done-for-you agent.
Your networks
Add a network
Or upload a data export
Your upload is read in memory and never stored. Account details are deleted after 24 hours; we keep only scores and your decisions.
| Account | Score | Why |
|---|
Removal jobs
Step-by-step removal instructions
Teach the agent
Show the done-for-you agent how you do it: a Bot Purge browser window opens, you do the steps once (clicks, typing, keyboard shortcuts, menus, new windows), then press Done in the bar at the top. Your steps are saved and the agent follows them from then on. Passwords and codes are never recorded.
While the agent works you can watch its cursor, press Take over (or just click or type in its window) to do something yourself, then Resume. If it can't find a button it asks you to click it, and remembers it next time.
Write or correct the steps
Platforms change their menus often. Write the steps that work for you; share them and a moderator will review before anyone else sees them. Never include links to login pages or ask for passwords.
Talk to your agent
Tell it how you want it to work, in your own words. It changes only your own settings and says exactly what changed.
When the agent doesn't know how
The agent knows the usual steps on each platform. When a platform works differently for you, or changes its menus, teach it one of three ways. It works for removing, blocking, reporting, unfriending and moderating your lives.
Show the agent: a Bot Purge window opens and you share your screen and cursor with the agent. Do it once while it watches, then press Done in its top bar. Passwords and codes are never recorded.
How the agent signs in
Sign in yourself (recommended): the agent opens the platform in its own window and you sign in there. Nothing is saved by Bot Purge.
Or save your sign-in: the agent signs in by itself. Only in the desktop app: it stays sealed on this computer and is typed only into the platform's own login page. If the platform asks for a code or a puzzle, the agent hands the window to you.
History
Removed someone by mistake? Open their profile, add them back, then press “I re-added them” so they're never flagged again.
| Date | Event | Account |
|---|
Scan your messages, comments and email
Bots and scammers show themselves by how they write and act: emoji codes like [tearsofjoy], pushing you to another app, the same line over and over, links, and scam pressure in email. We judge behaviour only, never anyone's opinions.
Gmail: takeout.google.com → Mail → .mbox. Outlook, Yahoo Mail, Apple Mail: drag messages out as .eml, or export a mailbox.
Android texts: the free "SMS Backup & Restore" app → back up → upload the .xml. iPhone texts: back up to your computer (Finder/iTunes), then upload sms.db (the desktop app finds it for you).
Connect your inbox
Read-only by default: Bot Purge reads your Inbox and Spam folders to find scams but can't send, delete or change anything. Protect checks them every day.
Paste comments or live chat
Data exports don't include comments on your posts or live chat. Copy them from the app and paste them here, one per line as handle: message.
Flagged senders
| Sender | Score | Why |
|---|
Connected apps
Follower-growth, auto-like and "who viewed my profile" apps are bots acting through your account. Upload the connected-apps file from your data export: X connected-application.js, or Facebook/Instagram "Apps and websites".
Canary traps: catch AI comment bots
AI bots read your post and follow any instruction in it. People skip nonsense. Put a canary instruction in your caption, a pinned comment or your video, then paste the comments here. Anyone who repeats the phrase is a confirmed AI bot. Bot Purge never posts or uploads anything for you.
Live Guard: bots removed from your live chat as they appear
Bot Purge moderates your stream's chat: fake giveaways ("first 10 to type WIN get $10,000"), "check my bio" bots, crypto doubling, impersonators of you or your mods, floods and scripted raids. It never touches you, your moderators, VIPs, subscribers or people you trust. It judges behaviour, never opinions. Start in Watch mode to see what it would do, then switch to Protect.
Your chat rules (optional)
Bots and scams are always caught. These rules are yours to switch on for your own chat, and they apply the same to everyone, whatever side they're on.
Connect your moderator account (Twitch / YouTube)
Twitch: make a moderator account (or use your own), create a token with moderator:manage:banned_users, moderator:manage:chat_messages and chat:read. YouTube: an OAuth token for the channel owner or a moderator with the youtube.force-ssl scope; the channel is the live chat ID.