WhatsApp chat analyser
WhatsApp chat analyser that quotes its evidence
Upload a WhatsApp export and get a plain-English report on what the conversation shows: the timeline, the patterns and the moments that matter, with the messages behind its findings quoted. Names in the chat text are replaced before the AI reads it.
- Findings quote the messages they rest on
- Names in the chat replaced before the AI reads it
- Refunded automatically if a report fails
- Built in Manchester by Debut Web Consultants
AI-assisted analysis of the messages you supply. It is not legal evidence and not an expert report. Only for conversations you were part of.
What the report gives you
A word count can tell you who sent more messages. It can't tell you what happened. Chat Forensics reads the conversation, writes up what it finds and quotes the messages its findings rest on, so you can check them rather than take them on trust.
- Quoted findings: short excerpts with the sender's label, the time and the message reference.
- How sure each reading is, in words (high, moderate or low), and why. Never a score.
- The other explanations each reading considered, and what would change it.
- Message counts, reply times and gaps measured from the export itself, not estimated by the AI.
- A sequence of events, transcripts of voice notes, descriptions of images, and a PDF of the whole report.
Ask the question you actually have
Each report type asks a different question of the same chat:
- Relationship analysis
- In-depth analysis of relationship dynamics and communication patterns
- Claim consistency
- Whether each person's account holds together over time, checked against what the chat itself records
- Denial and contradiction patterns
- Where one person denies, minimises or rewrites what the chat itself records, with both messages quoted
- Behavioural pattern detection
- Message frequency, response timing, avoidance and recurring interaction patterns
How it works
- Export the chat from WhatsApp: the .txt file, or the ZIP with the media in it. There is no WhatsApp login and nothing to install.
- Upload it. A short free preview runs straight away, so you can see the chat was read correctly.
- Pick a report type and check the itemised quote. It is measured from your upload and never goes up after you pay.
- Read the report in your dashboard. If a paid report can't be produced, you are refunded automatically.
What it won't do
It doesn't detect lies, diagnose anyone or score people or relationships. Every participant is read by the same rules, including you, and each finding says what the messages can't show. It never opens anyone's phone or account: you upload an export of a chat you were in. See exactly how it works
Questions people ask
- Is this an AI WhatsApp chat analyser?
- Yes. An AI model reads the chat text after participant names, phone numbers, email addresses and links have been replaced with labels such as Participant A. Voice notes and images you attach are sent as they are, because they can't be anonymised.
- Can it analyse a relationship chat?
- Yes. Relationship analysis looks at communication patterns and how the balance between you shifts, with quoted examples. It describes the messages. It doesn't rate the relationship or tell you what to do about it.
- Do I need to connect my WhatsApp account?
- No. Chat Forensics never connects to WhatsApp or touches your phone. You export the conversation with WhatsApp's own Export chat feature and upload the file in your browser.
- What does a report cost?
- The preview is free. Each report type then has a base price, and longer chats and attached voice notes and images add to it. You see an itemised quote before you pay, and it never goes up afterwards.
Start with a free preview
Upload the export and a short free preview runs first, so you can see the chat was read correctly. If you want the full report, you see an itemised quote before you pay. Reports start from $13.99 / £10.99. Get my free preview
The sample report is built from a made-up chat, so you can see how findings are set out before you upload your own. See a sample report