Maximize your WhatsApp conversations with WhatsWrapped insights

Getting Started with WhatsWrapped

There is nothing to install. WhatsWrapped is a web page: you export a chat out of WhatsApp, open whatswrapped.io, and drop the file onto the page. The first recap is free and includes every slide and every share image. After that a recap of a different chat costs about the price of a coffee, and sharing the images is never charged for on any tier.

You also do not configure anything. The deck you get depends on the file: a two person chat gets the pair deck, a group gets the group deck, and the two are genuinely different rather than one deck with slides hidden. With six people in a chat there is no "who is more into whom" to measure, so that slide does not exist in the group deck instead of existing and lying.

In the couple of seconds after you drop the file, the parser works out what it is holding. It samples the first few hundred lines to detect the platform it came from, the date order, the year width, the separator, whether the clock is twelve or twenty four hour, and the language. It commits to one reading, then scans the whole file for a date that contradicts it. That last step is what stops a chat with no day above the twelfth from being silently misread all year.

Understanding Your Conversation Data

The volume numbers come first and are the least surprising: messages, words and characters, how many days of the window you were actually active, and how many messages you send on a day you talk at all. Alongside them a timeline picks out your busiest single day and your busiest month, which is usually the slide people recognise an event in.

The split is the share of messages each person sent. On top of it sits an asymmetry score combining four measures rather than one: who starts the day, weighted heaviest at thirty percent, who double texts at twenty five, who writes more at twenty, and who replies faster at twenty five. Anything inside plus or minus fifteen hundredths is reported as dead even, because most chats are, and forcing a winner out of noise would make the number useless.

Reply time is deliberately two numbers, because the label covers two different questions. One is how long a message waits when it arrives cold. The other is how fast you volley when you are both already in the conversation. A silence of thirty minutes or more is treated as the end of a session, which is what keeps an overnight gap from being averaged in as a very slow reply and dragging the figure into meaninglessness.

The language slides are the ones people quote back. Your top five emoji, counted per person. Your top expressions, meaning the short phrases you actually repeat. And a signature word, which is not your most used word: it is the word you use far more than the other person does, relative to how much each of you writes, which is why it is usually a surprise rather than "the".

Then the shape of the whole thing: your longest streak of consecutive days, your longest silence and which of you broke it, and a closing receipt that itemises the lot on one page. A group deck adds a leaderboard of who carried it, a group wide median reply time, the roles people fall into, and the pair inside the group whose messages are mostly aimed at each other.

What is not in there is any reading of mood. No sentiment score, no emotional tone, no guess at how anybody felt. Every figure above is a count, a median or an ordering, and the deck states its limits where they matter rather than in a footnote.

Use Cases for WhatsWrapped

For couples the interesting slides are rarely the volume ones. Who starts the day, who sends the second message before the first was answered, and whether the reply gap between you is as lopsided as one of you claims. The asymmetry score exists for exactly this argument, which is also why it refuses to name a winner when the four measures disagree.

For long friendships the value is in the years. A streak of forty nine days and a silence of eight, with a name attached to who ended it, tells you more about how a friendship actually runs than any count of messages. Friends who drifted and came back can usually point at the month it happened on the timeline.

Long distance chats get the most out of the clock. Two people in different time zones produce a very readable pattern of when each of them is awake and writing, plus the latest message either has ever sent, which is often the thing that gets screenshotted.

Groups are a different question entirely. A leaderboard settles who carries the chat, the roles slide gives everyone a label they will argue about, and the closest pair usually confirms something the group already suspected. It works on family groups and friend groups alike, and the export button for a group is in the group info screen rather than the chat. What none of it settles is why, which is the standing limit on all of these numbers.

Privacy and Security Features

The chat file is read in your browser, parsed in a background worker on your own device, and never uploaded. There is no upload endpoint to send it to, which is a different and stronger claim than a promise not to look at what arrives.

Refreshing the page loses the chat, because it only ever existed in memory. That is deliberate. It is also the honest cost of the design, and worth knowing before you close a tab halfway through a deck.

Two things are stored, and the privacy page says so in the same words: your email address with how many recaps you have left, and a row recording that a recap ran and how many messages the chat contained. Not who was in it, not a word of what was said. A recap kept in your library stores the computed statistics, so opening it later replays the real deck rather than photographs of one.

One thing worth saying out loud: a group export contains other people’s messages. Nothing in the design changes that, and it is a reason to think about whose chat you are dropping in, not only about which tool you drop it into.

Advanced Features and Customization

The window is adjustable. The same chat can be measured across the current year, a rolling recent period, or its entire history, and switching between them changes every number on the page. A friendship of six years and the same friendship since January are genuinely different chats, and the year to date view is the one that reads like a wrapped.

Behind the deck sits a full numbers page, which lays out every statistic the engine produced rather than the dozen the slides had room for. Individual figures carry an explainer you can open in place, so a number you distrust can be traced to how it was defined without leaving the page.

Every slide can be exported as an image at the size a story wants, and the whole deck can be saved as a single PDF with one slide per page, which is easier to keep and to hand to somebody than a folder of files. Sharing an image is free on every tier.

Kept recaps go in a library, with five slots to start and the option to add more. Nothing is ever evicted to make room for something newer: a full library asks you to free a slot or add one, and leaves your existing recaps alone.

Troubleshooting and Support

If the page has gone blank and asks for a file again, you refreshed. The chat lives in memory only, so a reload genuinely loses it and the fix is to drop the file in again. This surprises people often enough that it is worth stating before it happens rather than after.

Lines reading "media omitted" are not an error and nothing is missing. A without media export replaces each photo, video or voice note with a placeholder, and those placeholders are what the media counts are built from. Exporting with media makes the file enormous and improves nothing in the deck.

If the dates look wrong, the likely cause is the day and month ambiguity that every export tool faces: 03/04 is either the third of April or the fourth of March, and only the file can say which. The parser resolves it by scanning the whole chat for a date with a day above twelve, so the failure case is a short chat that happens to contain none. A longer window usually fixes it.

A group whose Export chat button you cannot find is the commonest reported problem and it is not a limitation: on both iOS and Android that button sits in the group info screen, not in the chat. Very large exports are slower rather than unsupported, and choosing without media is the single biggest thing that helps.

For anything the deck itself does not explain, the in place explainers cover how individual numbers are defined, the glossary works through the terms one at a time, and the demo runs the whole thing on a made up chat with no account and nothing installed. There is no support desk to email, which is worth being straight about rather than implying one exists.

Comparative Competitor Review

The comparison worth making is between approaches rather than brand names, because the feature lists in this category change monthly and a specific claim about somebody else’s pricing is out of date by the time it is read. There are three real alternatives to a tool like this one, and each is better than it at something.

Reading the chat yourself is unbeatable for meaning and hopeless for scale. You will understand what a message meant in a way no counter can, and you will not get through fifteen thousand of them, and you will not notice that one of you starts sixty four percent of the days. Scrolling is a good way to relive a chat and a bad way to measure one.

A spreadsheet gives you total control and costs a weekend. Getting an export into columns means writing a parser for WhatsApp’s format, handling both date orders, deciding what counts as a reply, and picking a session boundary before a single number appears. It is genuinely the right answer if you want a metric nobody offers, and it is a lot of work to arrive at figures a page can give you in seconds.

Then there is the large group of tools that ask you to upload the export to a server. Many are good, and a server can do things a browser cannot. The tradeoff is not hidden: your file, containing other people’s messages, is now somewhere else, and you are trusting a policy rather than an architecture. If a tool asks for an upload, the questions are where it goes and how long it stays, and the answer should be easy to find.

The Spotify Wrapped comparison people reach for is close but not exact. Wrapped is assembled by a company from data it already holds about you; this is assembled by you, on your device, from a file you exported yourself, which is why it can cover a chat nobody but the two of you can see. All three comparisons have their own page if you want them in detail.