# wyper.io

> Privacy-first browser extensions. Two focused products: Social Cleanup (bulk delete
> your own posts, likes and followers on X and Instagram, running in your own browser
> session) and AI Fact-Check (a 1-10 truth score with real sources for any page, text
> or video, cross-checked by two independent AIs). One wyper.io Pro key unlocks premium
> in both tools. Nothing runs in the background; the tools act only when the user asks.

## Category
Social media cleanup / bulk delete browser extension. AI fact-checking browser
extension and web app.

## Products

### wyper.io Social Cleanup (Chrome and Firefox extension)
- Bulk deletes the user's own posts, replies, likes, comments and followers on
  X (Twitter) and Instagram; bulk unfollow; Instagram direct messages (whole chats, since 1.6.44).
- Filters: categories, and a keyword filter that can be inverted (delete only items
  containing a word, or keep only those).
- The deletion runs inside the user's own logged-in browser session. No password,
  no OAuth token, no API key is stored on any server. Posts and credentials never
  leave the device.
- What does leave the device: a licence check (key plus anonymous install ID) and,
  only when the user clicks publish, a content-free cleanup receipt.
- Free: 50 removals per day. Pro: unlimited.
- Disclosure: bulk deletion is against the terms of service of X and Instagram.
  The extension defaults to a conservative speed. Deletion is permanent.

### wyper.io AI Fact-Check (Chrome and Firefox extension + web app at https://wyper.io/app/)
- Select a claim, or point it at a page or video, and get a sourced verdict: a 1-10
  truth score with real, linkable sources, counterarguments and open questions.
- Standard check: "Dual-AI cross-check" of two independent AIs (Google Gemini and
  xAI Grok), showing disagreement instead of hiding it, plus a "Why this score" trail.
- Full Spectrum only: "Triple-AI reading" - three models on ONE evidence set. The
  third voice does not search; it receives exactly the sources the main model got and
  reads them independently, so a disagreement is a different reading of the same
  material rather than a different search result. This applies to Full Spectrum only,
  never to the standard check.
- "Full Spectrum" adds steelmanned counter-views, source diversity and open
  questions, kept evidence-based. A "Proof Library" (Pro) keeps a private
  archive of verifications (claim, verdict, sources, date).
- Video checks: YouTube (10 min free, 30 min licensed) and X videos. TikTok is
  caption-only. Instagram is not supported.
- The web app adds: photo scan with on-device OCR in 12 languages (the photo never
  leaves the device), 20-second audio capture (transcript only, audio not stored),
  and share-to-check on Android.
- Fact-Check is NOT a local tool: the text the user submits is sent for verification.
  Personal data (emails, phone numbers, card numbers) is stripped locally before
  any text is sent.
- Free: 5 checks per day, plus 3 Full Spectrum and 3 Dual-AI cross-checks per month.
  Pro: Dual-AI throughout, 80 Full Spectrum credits per month (an X deep-dive costs
  2 credits). Power: 200 credits per month.

## Published measurements (we are the primary source)
Raw data and method are published, including the results that do not flatter us.
- Quote checking, 633 real citations re-fetched and checked
  (https://wyper.io/quote-check-study): 73% verbatim on the page, 19% of pages not
  reachable for automated readers (403, PDF, paywall), 6.4% not findable at all.
- Source reach, 24 claims with the expected root named in advance
  (https://wyper.io/source-reach-study): 63% primary sources, 14% legacy media,
  1% fact-checkers, and all 6 non-English roots reached. No origin filter is applied:
  sources are never ranked by whether they are mainstream or official.
- Do two AI models agreeing mean anything? 30 claims, 150 frozen sources: Gemini and
  Grok agreed in 29 of 29, including deliberately contested claims. Two models trained
  on overlapping text are correlated, not independent, so "confirmed by 2 AIs" carries
  little information. A third model of different origin disagreed in 3 of 26, every
  time with a machine-verified quote. This argues against a claim we used to make.
- Can AI text be detected reliably? 240 texts measured
  (https://wyper.io/ai-text-study).
- Public dispute register (https://wyper.io/disputes): append-only and hash-chained,
  so entries cannot be silently edited or removed.

## How Social Cleanup differs from other cleanup tools
The structural difference is account access. Most cleanup services ask the user to
sign in with their X account and then act through that authorisation. Social Cleanup
asks for nothing: its entire browser permission set is storage and activeTab, and it
works inside the tab the user already has open, as the user. Nothing is granted, so
nothing has to be revoked afterwards, and there is no token to lose.
Honest trade-off: tools holding account access can filter by date, process an uploaded
archive, and keep deleting on a schedule while the user is away, because they run on
their own servers with the access the user granted. Social Cleanup can do none of those
three. The scheduling one is structural rather than unfinished: with no authorisation
held, work can only happen in a tab the user has open. No tab, no run. It filters by
keyword, not by date. Comparison with sources: https://wyper.io/tweetdelete-alternative

## Positioning
- Privacy-first, professional, and evidence-based. A verification and research tool
  that SUPPORTS the user's judgment; it does not replace it.
- Fact-Check's Power tier is aimed at high-stakes work: journalists, researchers,
  legal, human-rights, policy, medicine/science communication, OSINT, and education.

## Pricing
- Free: $0. Pro: $24/month or $252/year. Power: $49/month or $528/year.
- One key unlocks both tools, 5 device seats.
- Heavy "Full Spectrum" / full-length-video research is metered as credits with
  optional top-up packs; credits never expire. BYOK (bring your own AI key) is a
  Power add-on.

## AI Spotter (web app only)
- A PROVENANCE report for an image or a video, not a detector. It never outputs a
  percentage of how "AI-generated" something looks; that number is precisely the
  harm the tool exists to prevent.
- Layers: C2PA content credential (checked against 55 official trust anchors, and
  only a *trusted* signature counts), file metadata, reverse image search, earliest
  documented publication, and quote verification against the source page.
- Video: still frames are taken from the file and each is searched separately.
- Video caveat, and it is measured: a content credential lives in its own container
  box beside the picture data, so ANY repackaging drops it, including a lossless one.
  Platforms repackage on upload. On a video taken off a feed that layer will almost
  always find nothing, and that is the normal case rather than a suspicion. Check the
  original file if you can get it.
- Price: 1 credit per image, 4 per video. Paid plans only - this is the one part of
  Fact-Check with no free tier, because a provenance report calls three paid services
  on every run. The text fact-check stays free (5/day, no account).

## What it does NOT do
- No AI detector and no "percent AI" score, for image, video or text. AI-text
  detection was measured across 240 texts and deliberately not shipped.
- No export or archive of posts before deleting; use the platform's own archive
  download first if you want a backup.
- No date range filter in Social Cleanup; filtering is by category and keyword.
- No native App Store or Play Store app; the web app installs from the browser (PWA).
- No fact-checking of Instagram content; TikTok video analysis is caption-only.
- Fact-Check does not run locally and is never claimed to.
- No background activity; nothing runs unless the user starts it.

## Key pages
- https://wyper.io/ : overview and pricing
- https://wyper.io/cleanup.html : Social Cleanup product page
- https://wyper.io/fact-check.html : AI Fact-Check product page
- https://wyper.io/app/ : Fact-Check web app (PWA)
- https://addons.mozilla.org/firefox/addon/wyper-delete-tweets-posts/ : Social Cleanup on Firefox Add-ons (live since 2026-09-18)
- https://addons.mozilla.org/firefox/addon/wyper-ai-fact-check/ : AI Fact-Check on Firefox Add-ons (live since 2026-09-18)
- https://wyper.io/how-to-mass-delete-tweets.html : guide, mass deleting tweets
- https://wyper.io/how-to-bulk-delete-instagram-posts.html : guide, bulk deleting Instagram posts
- https://wyper.io/instagram-delete-dms : guide, deleting Instagram DMs in bulk
- https://wyper.io/how-to-fact-check-a-video-or-claim.html : guide, fact-checking claims and videos
- https://wyper.io/privacy.html : privacy policy
- https://wyper.io/terms.html : terms
- https://wyper.io/support.html : support, or support@wyper.io

## Social
- X: https://x.com/wyper_io
- YouTube: https://www.youtube.com/@wyper-io
- LinkedIn: https://www.linkedin.com/company/wyper-io
- Instagram: https://www.instagram.com/wyper.io/
- Telegram: https://t.me/wyper_io

- [TweetDelete alternative: an honest comparison](https://wyper.io/tweetdelete-alternative): Side-by-side with TweetDelete, including the three things TweetDelete does better (date filter, scheduled deletion, archive upload). Key difference: wyper needs no account authorisation, only the storage and activeTab browser permissions.

- [Bulk deleting posts: the questions people actually ask](https://wyper.io/cleanup-faq): 19 answers including the Nos (no date filter, no export, no archive import, no unattended runs), the free daily limit of 50, platform throttling, and why deletion is permanent.

- [How wyper Fact-Check is different](https://wyper.io/fact-check-different): the four claims we make and the measurement behind each, including the unflattering parts. Twelve independent passes on one claim (Booster), Triple-AI reading on Full Spectrum, 63 percent primary sources and 1 percent fact-checkers across 24 measured claims, 73.1 percent of 633 citations verbatim, and no AI-text detector because 240 measured texts showed it does not work.

<!-- pages:start -->
## Topic pages

- [How to check if an image is AI-generated](https://wyper.io/ai-image-check): We measured AI detection and refused to ship it. Here is the method that works instead: provenance, chronology, and the claim the image is carrying.
- [How to check if a video is a deepfake](https://wyper.io/deepfake-check): Detectors guess. Four checks that decide: origin, chronology, the claim itself, and who benefits. Plus why we refuse to ship a detector.
- [AI fact-checking for journalists and researchers](https://wyper.io/fact-checking-for-journalists): For work that gets published: verbatim quote verification, primary-source retrieval measured at 63 percent, counter-positions, public dispute register.
- [Fact-checking tools compared: which one for which job](https://wyper.io/fact-checking-tools-compared): Snopes, PolitiFact, Ground News, Google Fact Check Explorer and wyper do different jobs. Which to reach for, including when the answer is not us.
- [How to fact-check a health or pandemic claim](https://wyper.io/health-claim-check): From a viral post to the primary study: preprint versus peer review, absolute versus relative risk, and the four questions that settle most health claims.
- [Teaching fact-checking: three exercises that fit one lesson](https://wyper.io/media-literacy-classroom): Lateral reading and SIFT, three classroom-ready exercises, and the measurement showing why AI-text detectors should never be used to accuse a student.
- [How to verify war footage before you share it](https://wyper.io/verify-war-footage): Old clips, wrong war, real deaths. The verification order used by newsrooms, written for people who are not in one, and where automation stops.
<!-- pages:end -->

<!-- blog:start -->
## Blog

- [Why Do AI Models Disagree on a Fact-Check?](https://wyper.io/blog/why-ai-models-disagree) (2026-10-12): When AI models evaluate claims, they often reach different conclusions. We explain why training data, source selection, and reading methods cause friction.
- [Does deleting all my tweets delete my X account?](https://wyper.io/blog/delete-tweets-vs-account) (2026-10-11): Understand the difference between clearing your post history and deactivating your X profile, including the 30-day deletion window and what stays public.
- [How do I fact-check out-of-context videos on social media?](https://wyper.io/blog/check-out-of-context-video) (2026-10-10): Learn how to verify if a viral video is old footage relabelled as new. We explain how to check earliest appearance and event dates to spot false context.
- [Can I protect tweets from deletion?](https://wyper.io/blog/protect-tweets-from-deletion) (2026-10-09): You can protect important posts during a bulk cleanup by using a keyword filter with an invert option, keeping tweets that contain specific terms safe.
- [How do I check if a video has fake AI audio?](https://wyper.io/blog/check-fake-ai-audio) (2026-10-08): Learn how to verify viral videos with AI-cloned voices or recycled audio. We explain how to trace original footage and check the underlying claims.
- [Why does the tool only delete my newest tweets?](https://wyper.io/blog/newest-tweets-deleted-limit) (2026-10-07): The X platform caps timeline access at 3,200 posts. Learn why cleanup tools hit this API wall, how archive imports bypass it, and how to manage your history.
- [How do I verify a quote from a screenshot?](https://wyper.io/blog/how-to-verify-a-quote) (2026-10-06): Learn the three-step process to trace viral quotes back to their primary sources, check their original context and spot fake attributions online.
- [Bulk delete tweets privacy: do posts go to a server?](https://wyper.io/blog/bulk-delete-tweets-privacy) (2026-10-05): Find out if bulk deleting your tweets uploads them to a third-party server, and learn the difference between connected apps and browser extensions.
- [How do I check if a news broadcast video is fake?](https://wyper.io/blog/verify-fake-news-video) (2026-10-04): Learn how to verify viral news broadcasts and breaking video clips. We explain how to check video provenance, audio splicing, and AI-generated deepfakes.
- [How often should I clean up my social media?](https://wyper.io/blog/social-media-cleanup-schedule) (2026-10-03): A practical guide to building a realistic social media cleanup schedule. Learn what to review monthly and what to bulk-delete quarterly.
- [How to verify breaking news videos online?](https://wyper.io/blog/verify-breaking-news-videos) (2026-10-02): Learn how to check if a viral video is AI-generated or old footage recycled out of context, using the recent Flydubai flight incident as a guide.
- [How do I prove I deleted my posts?](https://wyper.io/blog/prove-deleted-posts-receipt) (2026-10-01): A deletion receipt proves you removed social media posts without revealing the deleted content. Learn how to generate an audit trail for your X and Instagram.
- [How do I fact-check a statistic?](https://wyper.io/blog/fact-check-a-statistic) (2026-09-30): A practical guide to tracing social media statistics back to their primary sources, verifying methodology, and identifying denominator errors in viral claims.
- [Should I delete my old posts or make my account private?](https://wyper.io/blog/delete-posts-or-go-private) (2026-09-29): Going private controls who sees your timeline today, while deleting removes the content itself. Learn the differences and what neither option can fix.
- [How to tell if a protest video is old footage?](https://wyper.io/blog/verify-old-protest-video) (2026-09-28): Old footage relabelled as new is the most common form of video misinformation. Learn how to verify the date and location of viral protest clips yourself.
- [How do I clean up my X account after a job change?](https://wyper.io/blog/clean-up-x-account-job-change) (2026-09-27): A practical guide to auditing and cleaning up your X history after changing jobs. Learn which categories to clear and how to use keyword filters effectively.
- [How to find the original context of a viral video clip?](https://wyper.io/blog/find-video-original-context) (2026-09-26): Discover how to trace a short video clip back to its source broadcast, verify its original date, and check if the footage has been deceptively repackaged.
- [How do I bulk delete Instagram comments?](https://wyper.io/blog/bulk-delete-instagram-comments) (2026-09-25): Learn how to manage and bulk delete Instagram comments using native Activity Center tools and browser extensions without giving up account access.
- [How to verify a screenshot online?](https://wyper.io/blog/verify-a-screenshot) (2026-09-24): Screenshots are easily faked and carry no inherent proof. Learn how to verify a screenshot using source checks, metadata and text fact-checking workflows.
- [How do I delete my old retweets and reposts?](https://wyper.io/blog/delete-old-retweets) (2026-09-23): A complete guide to finding and removing old reposts on X, why they are often forgotten during profile cleanups, and how to delete them in bulk.
- [How to fact-check a viral speech video online?](https://wyper.io/blog/fact-check-viral-speech-video) (2026-09-22): Learn how to verify altered political speeches and recycled video clips using provenance data, earliest appearance checks, and source tracking.
- [Can you undo delete tweets on X?](https://wyper.io/blog/undo-delete-tweets-x) (2026-09-21): Once a tweet is deleted from X, it is gone from their active servers. Learn why deletion is permanent and what to save before you clear your timeline.
- [How do I verify an altered video online?](https://wyper.io/blog/verify-altered-video-online) (2026-09-20): Learn how to check if a viral video is altered, why pixel detectors fail, and how tracing the earliest appearance uncovers the truth behind doctored clips.
- [How do I remove followers on X without blocking them?](https://wyper.io/blog/remove-followers-x-soft-block) (2026-09-19): Learn how to soft-block users on X, use the native follower removal feature, and clean up your follower list in bulk using browser-based tools.
- [How can I check if a photo is AI-generated?](https://wyper.io/blog/check-photo-ai-generated) (2026-10-02): Start with content credentials, then reverse-search for the earliest copy, then compare the claimed date and place. Steps that work, and what no tool can tell.
- [wyper is now on Firefox: what changed and what did not](https://wyper.io/blog/wyper-now-on-firefox) (2026-09-18): Both wyper extensions are live on Firefox Add-ons. How Firefox handles permissions differently, what we had to declare about data, and what stays the same.
- [How do I unfollow everyone on X at once?](https://wyper.io/blog/unfollow-everyone-x) (2026-09-17): X does not have a native mass unfollow button. Learn how to clean up your following list safely, understand platform limits, and avoid account locks.
- [How do I verify a deleted post screenshot?](https://wyper.io/blog/verify-deleted-post-screenshot) (2026-09-16): A practical guide to verifying screenshots of deleted social media posts using web archives, caches, and digital footprints to prove if a claim is real.
- [What is the best TweetDelete alternative in 2026?](https://wyper.io/blog/best-tweetdelete-alternative) (2026-09-15): Compare X cleanup tools safely. Learn how OAuth access, date filters, and keyword search separate cloud services from local browser extensions.
- [wyper Fact-Check update: Community Notes, video vs. text](https://wyper.io/blog/fact-check-update-community-notes-video) (2026-09-15): Three changes after one wrong verdict: the checker reads X Community Notes, tests whether a video shows what a post claims, and stops averaging unproven claims.
- [How do I check a claim in a language I do not speak?](https://wyper.io/blog/check-foreign-language-claim) (2026-09-14): Language barriers are the blind spot of online verification. Learn how to trace and verify non-English claims using AI tools and primary source evidence.
- [How many tweets can I delete per day?](https://wyper.io/blog/delete-tweets-rate-limit) (2026-09-13): X limits accounts to 2,400 posts per day, but rolling half-hour caps stop bulk deletion much sooner. Learn how to clean up your timeline safely.
- [Two vs three AI models for fact-checking?](https://wyper.io/blog/multiple-ai-fact-check) (2026-09-12): Why two AI models usually agree on a claim, and how adding a third model with different origins creates genuine contradiction in fact-checking.
- [Does Deleting a Tweet Remove It From Google?](https://wyper.io/blog/deleted-tweets-google-search) (2026-09-11): Deleting a post on X removes it from your profile, but Google caches often keep it visible. Learn how to clear deleted posts from search results entirely.
- [Why should I trust a fact checking tool?](https://wyper.io/blog/trust-fact-checking) (2026-09-10): Blind trust is a vulnerability. A reliable fact-checking tool provides raw data, verifiable evidence chains, and a public error rate you can check yourself.
- [Is using a tweet deleter safe?](https://wyper.io/blog/tweet-deleter-safe) (2026-09-20): We examine the security risks of OAuth tweet deleters, server-side data leaks, and how browser-based cleanup tools offer a safer alternative.
- [Is this image AI-generated? How to check a picture posted online](https://wyper.io/blog/check-ai-generated-image) (2026-10-02): No detector can tell you for sure. Three checks can: content credentials, reverse search for the earliest copy, and the context the post claims.
- [If I delete a tweet, are my replies and likes deleted too?](https://wyper.io/blog/what-happens-deleted-tweets) (2026-09-20): No. Deleting a tweet removes only that post. Your replies, reposts and likes stay on your profile. What survives, what does not, and how to clear the rest.
- [How do I verify breaking news claims in the first hour?](https://wyper.io/blog/verify-breaking-news-first-hour) (2026-09-06): The first hour of a news event carries the highest risk of false narratives. Learn how to verify origin accounts, timelines, and visual provenance.
- [How do I delete only tweets containing a certain word?](https://wyper.io/blog/delete-tweets-keyword) (2026-09-05): Learn how to filter and delete specific posts on X using advanced search operators or a browser extension with keyword and invert options.
- [Who decides what is true in a fact-check?](https://wyper.io/blog/who-decides-what-is-true) (2026-09-04): When you verify a claim online, the outcome depends on the sources used. See how measuring primary evidence replaces editorial bias with verifiable data.
- [Can I clean up Instagram and X with one tool?](https://wyper.io/blog/clean-up-instagram-x) (2026-09-03): Compare how to clean up Instagram and X using native tools, multi-platform schedulers, and browser extensions without giving away account access.
- [How do I fact-check a health claim?](https://wyper.io/blog/fact-check-health-claims) (2026-09-02): Learn how to trace medical advice from viral posts to primary studies, evaluate risks, and use automated tools to verify health information.
- [How do I bulk delete my Instagram likes?](https://wyper.io/blog/bulk-delete-instagram-likes) (2026-09-01): Learn how to find your hidden interaction history and delete Instagram likes in bulk using the official app or a browser extension.
- [How do I bulk delete my ChatGPT conversations?](https://wyper.io/blog/how-to-bulk-delete-chatgpt-conversations) (2026-08-31): ChatGPT has no native button to bulk-delete selected chats. Clear everything at once, delete one by one, or use a local browser extension.
- [How do you fact-check a post on X inside the thread?](https://wyper.io/blog/x-summon-fact-check-bot) (2026-08-31): Reply @Wyper_check under any post and a sourced verdict lands in the same thread in about 20 seconds. What it checks, what it refuses, and the costs.
- [How do I verify a war video before I share it?](https://wyper.io/blog/verify-war-video) (2026-08-29): Learn how to verify war videos using geolocation, timeline checks, and source analysis before you share them on social media.
- [How do I delete all my Instagram posts at once?](https://wyper.io/blog/delete-all-instagram-posts) (2026-08-28): A complete guide to clearing your Instagram grid using the native Activity Center and browser-based extensions for bulk deletion without API access.
- [How do I check if a video is a deepfake?](https://wyper.io/blog/how-to-detect-deepfake-video) (2026-08-27): Discover why pixel detectors fail to identify deepfakes reliably and how you can use context, origin tracing and claim verification to find the truth.
- [Do I need my Twitter archive to delete old tweets?](https://wyper.io/blog/delete-tweets-without-archive) (2026-08-26): You do not need to wait days for your X data file to clean up your profile. Here is how browser-based tools remove old posts directly from your timeline.
- [Is this viral video really from the event it claims?](https://wyper.io/blog/verify-viral-video-context) (2026-10-02): Often not. Old footage gets reframed as breaking news. How to find the first upload, compare dates and places, and read the Community Note first.
- [Can I delete tweets without giving an app access to my account?](https://wyper.io/blog/delete-tweets-without-access) (2026-10-02): Yes. A browser extension can delete your posts inside your own logged-in tab, without OAuth, API keys or your password. How it works and its limits.
- [How do I delete all my tweets at once?](https://wyper.io/blog/how-to-delete-all-tweets) (2026-08-23): X has no native button to bulk-delete posts. You must use third-party apps or browser extensions to clear your timeline. Deletion is always permanent.
- [How do you verify translated or subtitled videos?](https://wyper.io/blog/verify-translated-subtitled-videos) (2026-08-23): Learn how to check if a viral video's subtitles or dubbed audio match what the speaker actually said, using manual searches and AI verification tools.
- [How do I audit my social media footprint and clean it up?](https://wyper.io/blog/personal-privacy-audit-cleanup) (2026-10-02): List what a stranger can see, decide what stays, then bulk-delete the rest on X and Instagram. A practical audit in five steps, done locally.
- [How do you verify leaked texts and medical statistics?](https://wyper.io/blog/verify-leaked-texts-statistics) (2026-08-17): Learn how to trace viral screenshots of leaked documents back to their primary sources and verify distorted medical statistics using real evidence.
- [How do you clean up your social media for a job search?](https://wyper.io/blog/clean-social-media-job-search) (2026-08-13): A practical guide to auditing your digital footprint and pruning old social media posts, replies, and likes before your next job interview.
- [How do you spot AI-generated fake videos on social media?](https://wyper.io/blog/spot-ai-fake-videos) (2026-08-10): Learn how to verify viral social media videos by spotting AI visual glitches and contextual errors, using a recent fake London Tube video as a guide.
<!-- blog:end -->