The whole WhatsApp thread, read once and written down
Clark takes the .zip or .txt that WhatsApp exports — group chat or one-to-one, any language — and writes a report anchored to real lines from the thread. Photos, videos and voice notes never leave your device.
What you can upload
The .zip that WhatsApp produces from “Export chat”, or the .txt inside it. Either is accepted, and the upload step shows you the message count before you go any further.
Group chats and one-to-one chats, from three people up to a group of thirty. Chats in any language: the report is written in the language you choose, and every quote stays in the language it was written in.
Media is stripped in your browser before anything is sent, so a few hundred megabytes of photos become a text file of a few hundred kilobytes. The text of the conversation is uploaded, because that is what has to be read to write the report.
What Clark reads out of a WhatsApp export
A WhatsApp export carries more than the sentences. These are the things Clark reads for:
The timeline, not the average
Message counts by week rather than one total. Long silences and sudden busy stretches become part of the story instead of being flattened into an average.
Every participant, separately
Each person is counted and read on their own. In a group of ten, someone with forty messages in three years is a finding — and the report says the sample is thin rather than filling the gap.
Groups read as groups
A group chat is several relationships with an audience, not a pile of two-person conversations. Clark keeps track of who is talking to whom.
The lines each finding hangs on
Every chapter is anchored to quotes that are checked back against the file and rendered as chat bubbles in the report, in the language they were written in.
System noise, removed
“Messages and calls are end-to-end encrypted”, “You deleted this message”, joined and left notices, missed calls: none of that is conversation, and none of it reaches the report.
Why the whole thread, and not the last hundred messages
The last week of a chat tells you what happened last week. A year of it tells you who proposes, who answers, what repeats and which month the rhythm changed — none of which is visible in a sample.
WhatsApp exports are usually longer than people expect, and that is the point. A twenty-thousand-message thread is read as one conversation, in order, with a timeline, and the report is written from that.
Right after uploading you see the message count and the signal strength for that file. Under three hundred messages the report will be thin, and we show you that before you pay anything.
Questions people ask about this one
How do I export a WhatsApp chat?+
Open the conversation, choose Export chat, decide whether to include media, and keep the .zip or .txt the app hands you. Android does it from the chat menu, iPhone from the contact or group info screen. Both routes are written out step by step on the export guide linked at the bottom of this page.
Should I include media in the export?+
No. Choose “Without media”. Nothing Clark reads is inside a photo or a voice note, and including them can turn a small file into a very large one for no benefit.
Is my .zip uploaded?+
No. The archive is opened on your own device and the media inside it is discarded there. The text of the conversation is uploaded, and it is deleted on the schedule described in the safety notes.
Does it work on a group chat?+
Yes, and group chats are the case the report was built around. Every participant is read separately, and the group is treated as the several relationships it is.
My export is enormous. Is that a problem?+
Long threads are normal here. They are analysed in sections and then composed as one report with a single timeline, so length does not turn into a summary of a summary. What matters more than length is how much of the thread is about you: twelve thousand messages in a group where you sent two hundred is thinner than it looks.
Nothing is charged until you decide to read past the free part of a report.