wyper.io · hyper tools

Take back your digital life.

One privacy-first suite to clean it up and know what's true. Focused tools, one key, everything in your browser.

🔒 Privacy-first · runs in your browser 🔑 One key unlocks the whole suite
x.com / your posts
🧹 Cleaning your historyLIVE
Cleaning 0 / 1,2048,240 removed
🛡️ wyper.io · Verify
Analyzing…reading the claim…
checking
“…”
✓✓ Confirmed by 2 AIs (Gemini + Grok)
How it works

Three steps. Zero learning curve.

Clean your X, Instagram and ChatGPT history the same simple way - preview first, then sweep.

𝕏 X📸 Instagram🤖 ChatGPT
1
🎯

Pick a target

Posts, likes, replies and follows on X and Instagram, or your ChatGPT conversations. Filter by keyword so only what you choose goes.

2
👁️

Preview & confirm

See the exact items first - on ChatGPT the chats to be removed are marked and numbered, and pinned chats are kept. Nothing is deleted until you say go.

3
🧹

Watch it sweep

wyper works through the list in your own logged-in tab - steady, safe, with a live progress bar and mini-games for the wait.

1

Right-click anything

A page, a selected quote, or a video. wyper reads only what you point it at - nothing runs in the background.

2

Two AIs investigate

Gemini and Grok search the live web independently, then cross-check each other for a verdict you can trust.

3

Read the full picture

A 1–10 truth score, the finding in your language, sources - and with Pro, the counter-views and open questions.

Play while you wait

Cleaning years of history takes minutes.
So we made them fun.

A safe clean-up runs at a human pace - deleting thousands of posts can't be instant without risking your account. So instead of watching a progress bar crawl, drop into a quick built-in game. wyper keeps sweeping in the background; you look up, and it's done.

  • 🕹️ Real mini-games (Snake & more) - right in the extension
  • Turns unavoidable wait time into a 30-second break
  • 🧹 The clean-up never pauses - it runs while you play
  • 🙈 Not your thing? Hide the games in one click.
🎮 Play while you wait
🧹 Cleaning 823 / 1,204~2 min left

Live preview - the real thing plays inside the extension.

Why wyper

More than a tweet-deleter.

Depth other checkers skip.

🧹

Three surfaces, one tool

Most cleaners do X only. wyper handles X, Instagram and ChatGPT - posts, reels, highlights, likes, comments, story replies, and your AI conversations.

🎯

Selective by keyword

Don’t nuke everything. Delete only posts that match the words, topics or dates you choose. Keep the good, wipe the rest.

Follower control, both ways

Who you follow: unfollow everyone, or only people who don’t follow back.
Who follows you: remove followers, or clear non-mutuals. A smart follower scan runs first.

🔒

Your session, your data

Everything runs locally in your logged-in tab. No password, token or content is ever sent to a server.

Fact-Check

Searching the truth.

Not a one-word verdict from a black box. A full investigation you can read, cite and challenge - evidence first, no guessing.

8
Mostly accurate1 claim checked · cross-checked by 2 AIs
8/10 · supported- multiple independent sources agree
“Solar is now the cheapest source of new electricity in most of the world.”
Confirmed by IEA & IRENA cost data since 2023 - with regional exceptions where grids or policy lag.
Why this score
IEA/IRENA report solar LCOE below new coal in most markets
“Most of the world” hides regional exceptions (storage, grids)
✓✓ Confirmed by 2 AIs (Gemini + Grok)
🌐 FULL SPECTRUM CONTESTED EDGES
↔ Counter-view: critics argue “cheapest” ignores intermittency & system costs. Assessment: valid for full-system accounting; generation-cost claim still holds.
⚖ Disagree? Dispute this result with a counter-source - and it goes on the public log, whether we win or lose.
🤖
Two independent AIs, not oneGemini & Grok verify separately. When they disagree, you see it - no single black box.
🔗
Real, linkable sourcesEvery verdict is traceable to sources you can open and check yourself.
🚫
Evidence only - no guessingNo source, no verdict. It says “unverified” instead of inventing confidence.
🌐
Full Spectrum, unfilteredThe strongest counter-views, named and sourced - including voices beyond legacy media.
🎬
Pages, selections & full videosReads a whole YouTube transcript and returns the verdict in your language.
Contestable by anyoneDispute any published result with a counter-example. A test you can’t challenge isn’t verification.
🤖

Dual-AI cross-check

Gemini and Grok verify independently. Agreement gets a “✓✓ Confirmed by 2 AIs” - disagreement is flagged, never hidden.

🎬

Video fact-check

Checks the spoken words of a video in any language and returns the verdict in yours. Thai clip → German answer.

🌐

Full Spectrum Pro

The whole debate, not just the headline: the strongest form of every serious dissenting view - named, sourced, and honestly weighed - plus voices beyond legacy media. And a third model of different origin reads the very same sources, so agreement means something and disagreement is shown. Unfiltered, always evidence-based.

📚

Proof Library Pro

Every verification you run is saved to your private archive - claim, verdict, sources and date, viewable only with Pro, right inside the extension.

New · AI Spotter Web app only

Where did this image or video come from?

Drop in a screenshot, a photo or a video. You don't get a guess about how "AI-ish" it looks - you get a provenance report: what the file itself carries, where it has been published before, and who first documented it.

Try it in the web app
1
Content CredentialsA cryptographic signature inside the file. If Google, Adobe or a camera signed it, we verify that signature against 55 official trust anchors - and only a trusted one counts.
2
File metadataRead locally: dimensions, hash, and the IPTC marker that says "made by a trained model".
3
Reverse image searchWhere has this exact file appeared before? Full and partial matches, with the pages that carry them. For a video, every still frame we take is searched separately - a hit on any frame is a hit on the video.
4
Origin searchThe open web, for the earliest documented publication of the motif - and who published it.
5
Quote verificationThe sentence we cite is machine-checked, character for character, against the source text retrieved for that URL. What this does and does not prove.
Video works differently, and we say so. A content credential lives in its own box beside the picture data, so every upload strips it - even a lossless repackage. On a video that came off a platform, layer 1 will almost always find nothing, and that is the normal case, not a suspicion. The report says this instead of leaving you to guess. What carries the weight for video is layers 3 and 4: where these frames appeared first. That also happens to catch the most common real deception with video - old footage relabelled as something new - which no detector can see at all.
No score without evidence. If nothing is found, the report says "not determinable" and shows which layers came up empty. It never turns silence into a verdict - "nothing found" is not "nothing there". 1 credit per image · 4 per video · paid plans only
We measure our own chain and publish what it says. 63 % of the sources we surface are primary documents - courts, journals, datasets; legacy media 14 %, fact-checkers 1 %. 633 real citations re-fetched: 73.1% found verbatim on the page, and a quarter of the failures turned out to be our own fault. We also tested whether AI-written text can be detected, and it cannot - so we do not sell it: 240 texts, 720 raw judgements.
New · on 𝕏

Summon a fact-check anywhere on 𝕏.

See a claim in your feed? Reply to it with @Wyper_check and ask "is this true?". The bot reads the post you replied to, runs the same Full Spectrum check as the extension, and answers in the thread.

You
@you replying to a viral post
@Wyper_check is this true?
↓ about 20 seconds
W
@Wyper_check
7/10 · mostly accurate. Gemini and Grok agree on the core claim, disagree on the timeframe. Sources: iea.org, reuters.com. Full report with every source: wyper.io/p/…
Only when calledThe bot never posts uninvited. No mention, no reply - enforced in code, not by policy.
Opinions get refusedTaste, predictions and matters of opinion are declined, not guessed at. A checker that answers everything is worthless.
Every answer has a receiptThe thread reply links to the full report - claim, verdict, dissent and every source, open to anyone.
ContestableDisagree with a public result? Dispute it with a counter-source. Every challenge lands in a public, hash-chained log - open ones included.
See @Wyper_check on 𝕏
The consensus, the full debate, or the live open web.

How far do you want to see?

STANDARD CHECK · FREE

The mainstream verdict

A fast 1–10 truth score backed by the broad source consensus - one AI, one clear answer. Great for a quick, honest gut-check on everyday claims.

🌐 FULL SPECTRUM · PRO

The entire landscape

Triple-AI reading: three models, one evidence set. Two AIs cross-check every claim, and a third of entirely different origin reads the identical sources - we measured that the first two agree far more often than their independence would suggest, so the third is the one that can actually surprise you. The strongest version of each serious dissenting view - named, sourced, and weighed - with perspectives beyond legacy media and what's genuinely still contested. No filter, no bubble - but never false balance.

◆ DEEPEST
𝕏 FULL SPECTRUM + LIVE X-SEARCH · POWER

The fully open mode

Everything in Full Spectrum - plus a live pass across 𝕏 and the open web in real time. grok-4.5 with X-Search reads the signal as it breaks, before a consensus has even formed. The deepest look wyper can take: nothing filtered out, nothing waiting on the mainstream to catch up.

Power · for high-stakes work

Built for professionals who can’t afford to be wrong

When your work turns on the facts, the Power plan is your verification companion: dual-AI cross-checking, the full spectrum of views, primary documents and non-mainstream sources - and a Proof Library of every verification you can cite and revisit.

⚖️ Lawyers & legal teams 📰 Journalists & researchers 🕊️ Human-rights defenders 🏛️ Policy & public office 🩺 Medical & science communicators 🔍 Investigators & OSINT 🎓 Educators & academia

🔒 Privacy-first - local PII redaction, no data sale. A verification aid that surfaces evidence and dissent to support your professional judgement; it doesn’t replace it.

Compared

How wyper stacks up

FeatureTypical tweet-deleterswyper.io Cleanup
X (Twitter) cleanup
Instagram cleanup-
ChatGPT conversation cleanup-
Bulk delete ChatGPT chats (10 / 50 / 100 / all)-
Selective keyword deletelimited
Follower management (4 modes)-
Runs locally, no password sharedvaries
Live progress + mini-games-
FeatureSingle-AI checkerswyper.io Fact-Check
Live web-grounded verdict
Two independent AIs cross-checked-
Video / spoken-word checkingrare
Answer in your browser languagesometimes
Counter-views & open questions-✓ Pro
Private proof archive-✓ Pro
Pricing

Cleanup - clean without limits.

Fact-Check - go as deep as you need.

Each tool is its own extension with its own plan. One wyper.io Pro key unlocks the premium of every tool - you never pay per tool.

Free

$0

Up to 50 removals per day.

  • Bulk-delete posts, likes, replies, retweets
  • All follower & following modes
  • Keyword filters with invert
  • Capped at 50 removals/day (combined)
Add Cleanup - free
Pro

Cleanup - Pro

$24 / mo

Billed monthly · cancel anytime

  • Everything in Free, plus:
  • No daily limit - clean as much as you want
  • Priority processing & larger batches
  • Works on X, Instagram and ChatGPT
  • 🛡️ Fact-Check included - dual-AI truth scores with sources, same key
  • One key - every wyper.io tool
Get wyper.io Pro
🛡️
Your Pro key is also a Fact-Check key. The same subscription unlocks 🛡️ Fact-Check - verify any claim, page or video with dual-AI & real sources. One key, full power.
Explore Fact-Check →

Free

$0

Fast checks, everyday depth.

  • 5 fact-checks / day
  • Single-AI verdict (Gemini) + sources
  • Video checks: YouTube 10 min · 𝕏 clips 2 min
  • 3 Dual-AI cross-checks / month free
  • 3 🌐 Full Spectrum / month free - taste the premium
  • 📚 Proof Library
Add Fact-Check - free
Most popular

Fact-Check - Pro

$24 / mo

Billed monthly · cancel anytime

80 Full Spectrum deep-checks / mo
  • Unlimited standard fact-checks
  • Dual-AI cross-check (Gemini + Grok) · 80 checks / mo (then single-AI)
  • Videos: captioned to 2 h · AI-watched YouTube 30 min · 𝕏 7 min
  • 🌐 Full Spectrum (web) - counter-views & open questions
  • 𝕏 / social depth - Power, or 2 credits per dive
  • 📚 Proof Library - your private verification archive
  • 🧹 Social Cleanup included - bulk-delete on X, Instagram & ChatGPT, same key
  • One key - every wyper.io tool
Get wyper.io Pro

Power

$49 / mo

For professionals & high-volume research

200 Full Spectrum deep-checks / mo · incl. 𝕏 depth
  • Everything in Pro, plus:
  • 4× the deep-check allowance (166 / mo)
  • Dual-AI cross-check · 120 checks / mo (vs 80 on Pro)
  • 𝕏 / social depth included · 2 deep-checks per dive (≈ 100 dives / mo)
  • Video limits as Pro - 𝕏 over 7 min: coming to Power
  • 🔑 BYOK add-on eligible (coming soon)
Get Power

⚡ Top-ups - deepen whenever you need to

Out of Full Spectrum for the month? Add credits in one tap - any plan, no upgrade required. Top-up credits never expire (your monthly allowance refills each month). One web deep-dive = 1 credit · 𝕏 / social depth ≈ 5.

20 credits
$15
$0.75 / credit
Add 20
Most popular
60 credits
$39
$0.65 / credit · save 13%
Add 60
160 credits
$99
$0.62 / credit · best value
Add 160
🚀 Maximum · Power deep-search tool

Maximum Deep Dive

Twelve independent passes over one claim, each asking a different question: what the evidence says · who disputes it and why · what the primary documents record · the numbers · the chronology · sources outside the English-language mainstream · what would falsify it · who benefits · which adjacent claims it gets confused with · where the claim came from and how it spread · whether the position has since moved · and how the magnitude compares to the base rate. Plus a separate multi-hop deep-research run with its own citations. Different questions reach different sources. Measured on a real run: 24 counter-positions instead of 3, 48 sources instead of 8, 36 statements of which 31 carry a verified quote instead of ~5, and 12 open questions. All twelve run in parallel, so it finishes in about 70 seconds.

What 24 credits buy: pick one. 1 Booster: one claim, twelve questions. · 12 𝕏 dives: twelve claims, one question each. · 24 web deepenings: twelve levels down a single chain.
It is the same amount of research either way. The Booster bundles it onto one question; the dives spread it across twelve. Depth or breadth: the tool does not decide that for you.
Paid from purchased credits only, on every plan, including Power. A monthly allowance never funds a Booster, so nothing you already paid for gets used up by one.

  • 🛰️ Uncapped depth - no token ceiling
  • 𝕏 Live social signal + open web
  • 📑 Primary sources & minority views
  • 🎯 One claim, maximum rigor
Get Booster · $15 = 24 credits on your key · opt-in, one tap from any result · you see the cost before you run it
🔄 Auto-refill COMING SOON - top up automatically when you hit zero. Not available yet.
🎫 Day Pass, $3.99 - 24 h of full Pro access, no subscription. Key arrives instantly. Get a Day Pass →
🔑 Power BYOK COMING SOON - bring your own Gemini / Grok key for unlimited Full Spectrum at a flat +$9/mo. Built for the heaviest users. Not purchasable yet.

1 credit = one claim researched in depth (Full Spectrum, web) or a captioned video check. An 𝕏 / social deep-dive of a claim costs 2: on Power it comes out of the monthly allowance (≈ 100 𝕏 dives), on Pro out of purchased top-up credits. The opt-in 🚀 Maximum Booster costs 24 credits, and always from purchased top-up credits, on every plan, including Power. Your monthly allowance does not fund it. Those 24 credits are a choice, so in plain terms: one claim asked twelve different ways, or twelve separate 𝕏 dives on twelve different claims, or twenty-four web deepenings (twelve levels down a chain). Same credits, same amount of research, bundled onto one question instead of spread across twelve. Take the Booster when one claim decides something; take the dives when you are mapping a field. Standard fact-checks stay unlimited on Pro & Power. Monthly allowance refills each month; purchased top-up credits never expire.

🧹
Same key, second tool. wyper.io Pro also unlocks 🧹 Social Cleanup - bulk-clean your X, Instagram & ChatGPT history, locally in your browser. Full power, one subscription.
Explore Social Cleanup →
🔑 wyper.io Pro is one subscription & one licence key - it unlocks premium in both tools. Prices indicative for launch.
Good to know

Frequently asked

Do you ever get my password?
Never. Everything runs inside your own logged-in browser tab. No password, session token or private content is ever sent to a wyper.io server.
Is deleting posts safe for my account?
wyper works through your actions at a steady, respectful pace inside your session and shows a live count so you stay in control. You always preview and confirm before anything is removed.
Can I bulk delete my ChatGPT conversations?
Yes. Social Cleanup also clears ChatGPT: pick 10, 50, 100 or all of your own conversations. The chats to be removed are marked and numbered first so you can check, and pinned chats are always kept. It runs in your own logged-in ChatGPT tab with no OAuth, and it is not an official OpenAI API, so automated deleting is at your own account risk. Deletion is permanent. Read the guide.
What does the Proof Library store, and why is it Pro?
Every fact-check you run is saved as a permanent record - claim, verdict, sources and date. It’s your personal evidence archive, viewable only inside the extension with an active Pro licence. Without Pro, past verifications aren’t accessible.
Is one Pro plan really enough for all tools?
Yes. wyper.io Pro is a single subscription and licence key that unlocks premium across Cleanup, Fact-Check and every future tool - no per-tool checkout.
What’s a “deep-check”, and what happens when I run out?
A deep-check is the research-heavy work: one Full Spectrum analysis or a captioned video check. What the standard check is, plainly: a fast verdict that shows sources when it finds them, because we do not force a web search, because forcing one would cost more than the standard check is worth. Measured on 40 claims: it searches in about a third of cases, so two out of three standard checks come back without a source to click. That is the trade: standard is the quick look. Full Spectrum and 𝕏 depth always retrieve sources: ten of them per claim, and every statement they produce is either backed by a verbatim quote we located in that source, or labelled as coming from the AI archive with no source behind it. That is what you are paying for. An 𝕏 / social deep-dive costs 2 deep-checks; the 🚀 Maximum Booster costs 24 top-up credits. On Pro, cleanup and standard fact-checks are unlimited; the free plan includes 50 cleanup removals and 5 fact-checks per day. The Dual-AI cross-check is a separate monthly allowance - 80 claim cross-checks on Pro, 120 on Power, 3 on free. Beyond it a check still runs, on a single model (Gemini) instead of two, and nothing auto-charges. Your remaining Dual-AI allowance is shown in the plan overview in the app and the extension. When your monthly deep-check allowance is used up, nothing auto-charges: you simply choose to buy a top-up pack (20 for $15, 60 for $39, or 160 for $99) if you want more. Heavy users pay for what they run, so it stays fair for everyone else.
Can I use my own API key (BYOK)?
Soon. BYOK is a planned Power add-on (coming shortly after launch): connect your own Gemini or Grok key and run unlimited Full Spectrum for a flat +$9/month - ideal for very high volumes of deep research. Not purchasable yet.
Which languages does the fact-checker support?
It reads the source in its own language but always returns the verdict in your browser’s language - a Thai video gets a German answer for a German browser, correctly handling local calendars and conventions.
What is @Wyper_check on X (Twitter)?
It is our summon bot. Reply under any X post with @Wyper_check and it fact-checks the claim in the post it is replying to, then answers in the same thread with a verdict and any sources it finds. It only responds where it is explicitly mentioned - never on its own, and never anywhere it was not summoned.
Take control

Control your footprint. Verify the facts.

Two focused tools, one trusted brand - privacy-first, built for work that matters.

🚀 Pre-launch - get one email the moment wyper.io goes live. No spam, unsubscribe anytime.
- or join the community -
💬 Join our Telegram

People we work with