How to make a fake screenshot for every social platform — one tool, 26 UIs

A working fake-screenshot pipeline has two competing requirements. You want a single workspace, because every time you change tools you lose 90 seconds context-switching, hunting your character's avatar file, re-typing the contact name, re-picking dark mode. You also want pixel-accuracy to each individual platform, because the entire point of the screenshot is that it looks indistinguishable from a real one — generic "chat bubble" tools that try to do all platforms with one shared renderer always end up looking like none of them.

HeyFake's answer is to split the difference: every platform gets its own dedicated, pixel-accurate generator with platform-specific state (read receipts on WhatsApp, verified badges on X, gradient rings on Tinder), and a single workspace called the fake screenshot Studio that lets you switch between them from a sidebar without losing your zoom, your character or your tab order. The 26 dedicated pages still exist — they each rank for their own keywords, host their own how-to, and are the right place to land if you only ever make one type of fake.

Studio is the workspace; the dedicated pages are the landing pages. If you're making a single fake WhatsApp screenshot for a meme, the dedicated WhatsApp page is fine. If you're building a TikTok carousel that crosses four platforms in two minutes, the Studio is the difference between "ship in 10 minutes" and "ship in an hour".

Chat UIs are the bulk of the fake-screenshot genre — TikTok storytime, journalism illustrations, app mockups, design pitches. They all look superficially similar (avatars, bubbles, timestamps) but the detail that gives away a fake is different for every messenger. Going through them family by family.

WhatsApp (chat, chats list, contact info, calls)

WhatsApp is the world's messenger, which means everyone has used it and everyone's eye has been trained on its exact UI. That's a problem and an opportunity: a believable fake reads as effortlessly real, and a sloppy one is spotted in a second. The big tells in the genre are read receipts (two blue ticks = read, two grey = delivered, one grey = sent) being wrong for the storyline, the status-bar time being stuck at 9:41, and timestamps appearing on every single bubble instead of inline every few minutes.

The platform variant matters more than people realise. iOS WhatsApp bubbles are slightly more rounded; Android is tighter, and timestamp placement is different. If your character's name is "Carlos" and the bubbles look like iPhone WhatsApp, the cognitive dissonance is real even if the viewer can't articulate why. Most of the world uses WhatsApp on Android — picking the right variant is a free believability boost.

The HeyFake bundle covers four WhatsApp screens — the main chat, the inbox/chats list, the contact info profile, and the calls screen — which is enough to mock up an entire interaction arc, not just a single conversation. For the long version of how to make a single fake WhatsApp screenshot read as native, our deeper how-to guide covers contact-name choice, photo, pacing and read-receipt strategy in detail.

iMessage

iMessage's UI is the second most-instantly-recognised chat UI on earth. The tells are subtle but unforgiving: the blue (sender) bubbles have a specific tail shape that bends in a curve, the grey (receiver) bubbles use a softer rounded grey, and there is a small "Read" indicator under the last delivered message that appears nowhere else in messaging. The contact header always shows the name centered with a tiny "iMessage" subtitle directly beneath it — miss that subtitle and the screenshot reads as a generic chat tool, not iMessage.

Dark mode is huge for iMessage screenshots. Real iPhones default to dark mode for a large slice of users now, and a dark-mode iMessage screenshot in a TikTok video feels more "phone caught in the wild" than the brighter white-background variant. Pick one mode and commit — mixed-mode chats across multiple frames in the same video stick out.

iMessage is also where group-name conventions matter. Group chats in iOS often have a name like "Mom, Dad, Sis" or an actual title like "Friday plans" — the lazy fakes use a name like "Group Chat 1" and that single decision destroys the realism faster than anything else in the image.

Instagram DM

Instagram DM is the platform-of-choice for celebrity-style storytime fakes — the "Zuck DMed me $1 billion to buy my app" genre. The UI tells here are the purple-to-blue gradient on the sender (you) bubbles, the gradient ring around the recipient's avatar (story ring), the small "Active now" status under their username, and the camera button on the left of the input bar. Get any one of those wrong and the screenshot reads as a knock-off.

The little heart icon at the right of the input is a giveaway detail people often forget. Instagram added it to encourage one-tap reactions to the last received message; it's a UI element nobody mentions but everyone's eye has parsed. Replacing it with the wrong icon — or removing it entirely — is one of those instant tells that make even a well-written fake feel off.

Tinder Chat

Tinder is unusual: the chat itself is preceded by a "You matched on [date]" banner with a circular avatar in a pink-orange gradient ring. Skip that banner and the screenshot stops looking like Tinder. The bubbles are a blue gradient on the sender side and pure grey on the recipient — there is no read indicator under the bubble (Tinder doesn't do read receipts in the same way), and the input bar has a distinctive GIF button on the left that the other dating apps don't share.

Tinder fakes work best for storytime content where the entire arc is one match — first message to punchline. Cross-platform Tinder storytelling is rare because most viewers' mental model of Tinder is "here's where we matched, here's how it went", not extended threading. Keep the message count low (4-10 bubbles) and the punchline tight; long Tinder threads in screenshots feel artificial.

Discord Chat

Discord is the one chat UI that most heavily uses markdown — bold, italics, code blocks, spoilers, blockquotes — and any fake that ignores that ends up reading as generic chat instead of Discord. The other defining detail is the avatar layout: Discord stacks consecutive messages from the same author under a single avatar header, rather than repeating the avatar on every bubble like WhatsApp or iMessage. Miss that grouping and the rhythm of the screenshot reads wrong.

@mentions render in a soft blue tint, and channel mentions render the same way. Bot messages have a small "BOT" tag next to the username. Real Discord servers have a left-rail of channels and a right-rail of members — a fake that includes just the central chat column is fine for content; one that tries to also fake the rails has to get a lot more right.

Telegram (chat, channel, contacts, chats list, profile)

Telegram is its own continent in the fake-screenshot world because it covers five distinct screens that all matter. The main chat looks superficially similar to WhatsApp but with slightly different bubble shapes and a sender-on-left/receiver-on-right convention that often catches new fakers out. Channels are a different beast entirely — large header, subscriber count, broadcast-style posts with views counts rather than read receipts — and they're the heart of the Telegram-influencer culture in Russian, Persian, Turkish and Spanish-speaking markets.

The other three Telegram screens — contacts list, chats list and profile — are the ones nobody else fakes. Contacts shows your phone book with online status; chats list is your inbox; profile is a contact's page with phone, bio, media tabs. Together they let you mock up an entire Telegram session, not just a conversation snippet. If your story needs "they have 50,000 followers on a Telegram channel and slid into my DMs", the Telegram bundle is the only tool that lets you do both in one consistent visual style.

Social-post fakes are the second biggest genre after chats. They're used for satire, hot-take reaction memes, "screenshot of a tweet"-style content, journalism illustrations and brand mockups. The difficulty is that every platform's post UI is heavily redesigned every 18 months, and a fake that uses the 2022 version of a UI is instantly dated.

X (Twitter) Post

X is the most-faked social UI on the internet. The current 2026 version has a few specific tells: the verified blue badge means "subscribed to X Premium", not "verified person", and that change in meaning is part of the joke for most modern X fakes. The engagement bar shows replies, reposts, likes and views — views is the newer addition, and a fake that omits it dates itself instantly. Time format is relative for recent posts ("2h", "5m") and absolute for older ones ("Apr 22").

Card-shaped X fakes work best at desktop aspect ratio, not phone-shape — they're for sharing in articles, slide decks and Reddit posts where viewers expect to see a screenshot of someone's computer browser, not their phone. Match the aspect ratio to the use case.

Facebook Post

Facebook's post UI is where most fakes feel dated, because Facebook itself feels dated to anyone under 30. The current 2026 version has slimmer cards, a redesigned reaction picker (Like / Love / Care / Haha / Wow / Sad / Angry as standalone reactions, not just "Like"), and a clean comment-count + share-count row at the bottom. Older Facebook fakes still floating around the internet use the pre-2023 chunky borders — those read as nostalgic now, which works for "remember when" content but not for present-tense screenshots.

Date format on Facebook is unusually verbose — "Yesterday at 3:14 PM", "May 8 at 9:42 AM", or "Just now". Posts have a privacy icon next to the timestamp (globe for public, two-people silhouette for friends). Missing those is the difference between "Facebook post" and "generic social card".

Instagram Post

Instagram post UI has stayed remarkably stable since 2020 — square or 4:5 image area, a like-comment-share-bookmark row underneath, a caption that auto-truncates with a "... more" link, and a comment count line above the timestamp. The tells are subtle: the heart icon is outlined when unliked and red-filled when liked; the comment icon is a chat-bubble outline; the share icon is a paper-plane.

Caption format is a giveaway. Real Instagram captions are heavy on emoji, hashtags clustered at the end, line breaks between sentences. Fakes that use a single paragraph of clean prose read as press releases, not posts. The like count format is also specific: under 999 shows the number, then it switches to "1.2K", "15K", "1.2M" with no decimals after the millions.

LinkedIn Post

LinkedIn fakes are their own meme genre — the "motivational post about how I fired my entire team on Christmas Eve" style. The UI tells are corporate-blue accents, the connection-level badge next to the author's name (1st, 2nd, 3rd), the comma-formatted reaction count and the slightly-too-professional avatar treatment.

The crucial detail is the title-and-company line under the author's name — "Founder & CEO at [Stealth Startup]", "Building [vague AI thing]", "Helping leaders unlock their potential". The title format is what makes a LinkedIn fake unmistakeably LinkedIn. Posts often start with a single emoji or em-dash followed by a hook line; that opening rhythm is its own meme now and the easiest way to make a fake feel native.

TikTok Post

TikTok screenshots are weird because they're mostly screenshots of a video preview or a comment, not a static post in the way Instagram is static. The fakable surfaces are the video page with its right-side rail of like/comment/share/bookmark counters and music attribution at the bottom, and the comment thread overlay with avatar + handle + comment + reply counts.

Music attribution is where most fakes fail. Real TikToks show a small "♫ Original sound — username" or "♫ Song name — Artist" running ticker at the bottom; fakes leave it blank and the video page suddenly looks unmoored from the platform. The vertical aspect ratio is non-negotiable — TikTok fakes that aren't 9:16 read as "screenshot of a screenshot".

Threads Post

Threads has the cleanest, most-minimal social post UI in the cluster — by design, since Meta launched it as the calm-alternative-to-X. The card has the author's name, the verified badge if applicable, the post text, and a small Like / Reply / Repost / Share row underneath. No view counts (yet), no quote-post by default, no algorithmic engagement bar.

Threads fakes work for parodying the "earnest poster" aesthetic — long-form mini-essays in the post itself, replies that chain into the original. The chain rendering is the tell: real Threads posts show a small avatar stack on the left side of the card indicating a reply thread; missing that and the screenshot reads as a one-off note instead of part of a conversation.

Bluesky Post

Bluesky is the "the indie X" of 2026 — smaller, federated, less algorithmic. The UI tell is the handle format: every account is something dot bsky dot social or a custom domain. A real handle is unmistakeably AT-Protocol-style. Fakes that use plain @username read as wrong.

The engagement row is similar to early Twitter (replies, reposts, likes) without view counts or algorithmic boosts. Bluesky fakes are useful for tech-Twitter culture content, journalism reactions and any "the conversation that's happening outside of X" framing.

Pinterest Pin

Pinterest is the odd one out — image-first vertical cards with a pinner name, board name and red Pinterest logo. Fakes are useful for design mockups, mood-board content and parodies of the "rich aesthetic" genre. The tell is the vertical aspect ratio (Pinterest cards are taller than wide) and the "Saved by X people" counter at the bottom.

Pinterest's save-button is the second tell — a small red oval with "Save" on the upper right of the card hover state. A static Pinterest fake without that button reads as a generic vertical image; with it, it reads as the platform.

AI-conversation fakes are the newest big genre — the "ChatGPT said this insane thing" meme format that dominates 2024-2026 social. The challenge is that all the AI chat UIs look superficially similar (user message above, AI response below, model logo somewhere), but each has its own specific branding that's instantly recognisable to anyone who actually uses these tools.

ChatGPT Conversation

ChatGPT's UI is the one everyone's eye is trained on. The current 2026 version has the user message in a soft grey bubble on the right, the assistant response on the left without a bubble (just left-aligned text under a small OpenAI logo), and a thin model-name label at the top ("ChatGPT 5", "GPT-4o", "GPT-4o mini") that's the single biggest believability lever.

Get the model name wrong — say "ChatGPT 4.5" when no such public model exists at the time of the screenshot — and the fake reads as fake to anyone who follows AI. The other tells are the regenerate / copy / thumbs-up-down buttons that float under the AI response, and the small disclaimer line at the bottom ("ChatGPT can make mistakes..."). Skip the disclaimer and it doesn't feel like ChatGPT.

Claude Conversation

Claude (Anthropic) has a deliberately warmer aesthetic — soft cream background, orange accent on the Claude logo, slightly more typographic-feeling response styling. The user message is on the right in a small bubble; the Claude response is on the left with a hand-drawn-style sparkle logo. Compared to ChatGPT, Claude's output usually feels longer and more essay-formatted by default, and a fake that has a one-line Claude response reads as wrong.

Model names matter the same way they do for ChatGPT. "Claude Opus 4.7", "Claude Sonnet 4.6", "Claude Haiku 4.5" are the current public model line; using an older or invented name dates the fake or reads as obvious fiction.

Gemini Conversation

Gemini (Google) is the most-blue of the AI chat UIs — Google's blue runs through the model name, the send button and the small "Gemini" badge at the top. User messages are right-aligned with no bubble, just plain text on the page; assistant responses are left-aligned under a small four-pointed-star sparkle in Google's gradient blue.

Gemini's tells are the soft Google Sans typeface (different from ChatGPT's default), the "Show drafts" button under some responses (a Gemini-specific feature that lets you see alternate generations), and the rounded-square avatar shape that's subtly different from Claude's and ChatGPT's circles.

Grok Conversation

Grok (xAI) has the most distinctive aesthetic of the cluster — dark by default, X-platform-style typography, with an irreverent tone in the model's default responses. The current Grok UI has the user message on the right in a small dark bubble and Grok's response on the left under a small X-style sparkle logo. The platform context matters: most Grok fakes are screenshots of Grok inside the X app, not the standalone grok.com web UI.

Grok-3 was the most-faked version for memes in early 2026. The model-name label at the top, plus the "Powered by xAI" or X-app context, is the giveaway. Grok fakes that look like ChatGPT but with the word "Grok" pasted in read as obvious.

Perplexity Conversation

Perplexity is the odd one in the cluster — it's less a chat UI and more an answer-engine UI with inline source citations. The signature look is the question at the top, a structured answer underneath, and a row of numbered source cards (1, 2, 3, 4...) that the user can click to verify each claim. A fake Perplexity response without source cards reads as ChatGPT-in-Perplexity-skin, not actual Perplexity.

Perplexity fakes work for content where the joke is the citation — "Perplexity citing The Onion as a source for X"-style memes — or for hot-take posts illustrating an AI search result. The structured-answer format (bold subheadings, bullet lists, "In summary..." closer) is the second biggest tell after sources.

GitHub activity fakes are a small but specific genre — the contribution heatmap (the green-square grid showing daily commit counts for the year) is the meme target, used for "I coded every day for a year and built X" storytime or its inverse "every square is empty, here's why". The visual is instantly recognisable to developers and reads as "developer cred" to non-developers.

The believability factor is the cluster distribution. Real contribution graphs have spiky weeks, fully-empty weekends for normal-jobbed developers, holiday gaps in late December, and one absurdly green stretch where the developer was building something specific. A perfectly uniform green grid reads as fake; a realistic-looking one needs deliberate gaps. The other tells are the year label at the top, the streak counter (longest streak, current streak) and the daily-contribution-count line that hovers on tile mouseover.

Pulling together a multi-platform fake-screenshot piece in one Studio session looks like this. You open the page, pick the first platform from the left sidebar (let's say WhatsApp Chat), type your conversation, drag the zoom slider until the phone preview fits your laptop screen, click Download, get a clean PNG. You move to the second platform (let's say iMessage). The sidebar stays open, the zoom you set stays where it was, and you start typing the next scene — no page reload, no losing context, no "wait, what was the contact name I used three minutes ago".

The third scene is on Instagram DM. You click into the sidebar, scroll to the Social cluster — actually no, IG DM is a chat, so it's in the Chats cluster — click it, the editor on the left swaps and the preview on the right rerenders. Your zoom is still 80%. The character's avatar isn't carried over (each generator manages its own state, deliberately, so cross-platform state doesn't leak in surprising ways), but you've got the file in your downloads folder ready to upload.

The fourth scene is an X post. Card-shape, not phone-shape, so the zoom slider behaves differently — you scale to fit the card in your screen rather than the phone. Click Download. Five minutes later, four scenes, four downloads. The same workflow run across four dedicated generator pages takes 15-20 minutes and requires you to mentally context-switch four times.

Beyond the per-platform tells, there are a handful of universal mistakes that turn any fake screenshot into an obvious one, regardless of which UI you're forging. After hundreds of TikTok storytime carousels and our own listicle round-ups in the best fake screenshot generators comparison, here's the cross-platform list that we keep flagging:

1. Status-bar time of 9:41. The Apple keynote time. It's a tell on every iOS-style screenshot ever made. Change it to anything plausible for the storyline — 10:31, 2:14 PM, whatever.
2. Default silhouette avatar on the contact. The grey person-outline placeholder reads as "I didn't bother". Even a stock-art headshot, a blurred selfie or an emoji on a coloured circle is better than the default.
3. Wrong platform variant for the character. American character with Android WhatsApp. Brazilian character with iOS iMessage (iMessage doesn't exist on Android — that's an instant tell). Pick the platform that matches the character's background or the screenshot reads as wrong even when nothing specific is off.
4. Perfect grammar in casual chats. Real chat messages have lowercase first letters, missing punctuation, the occasional typo, abbreviated words. A WhatsApp screenshot where every bubble is correctly punctuated reads as a transcript, not a chat.
5. Timestamps on every single bubble. Most chat UIs collapse consecutive same-author bubbles under a single timestamp. Fakes that show a timestamp on every line look like log files.
6. Date format that doesn't match the locale. American MM/DD/YYYY on a screenshot from a Brazilian character. 24-hour times on an Apple-default screenshot. These break the realism quietly.
7. Stretched or squashed aspect ratio. Phone-shape screenshots that aren't 9:16. Card-shape screenshots that aren't the platform's native ratio. Both read as "screenshot of a screenshot" the moment they hit the timeline.
8. Mixed dark/light mode across frames. A storytime that switches between dark-mode and light-mode screenshots between scenes pulls the viewer out instantly. Pick one mode for the entire piece.
9. Watermarks from cheap tools. A bottom-right "Made with [tool]" tag is a sign the maker used a free-tier of a paywalled tool, and viewers who have used that tool spot it immediately. A clean screenshot has no watermarks.

Pass all nine of those and your fake reads as native on any platform. The platform-specific details still matter (Step 4 of our WhatsApp deep-dive walks through them for that platform specifically), but the cross-platform list is the floor.