About DryText

What DryText is

DryText (drytext.app) is a web app that analyzes exported WhatsApp and iMessage chats for dry-texting patterns — short replies, reply-time changes and who starts conversations — and links each observation to message excerpts.

People often arrive with a narrow question: does this exchange feel different, who keeps reopening a conversation, or are short replies usual for this chat? DryText is designed to organize the visible evidence so a person can review it more calmly. The app does not claim that a message can reveal another person’s private thoughts.

How a report is made

The workflow starts with a chat export that the user chooses. A WhatsApp TXT or ZIP export, or an iMessage TXT export, is parsed into dated messages and participant labels. ZIP extraction happens in the browser. The transcript is then uploaded to create the requested report; it is not a local-only analysis.

The report looks at observable signals such as reply length, changes in response timing, who starts conversations, and how each person participates. An AI model helps explain those patterns in readable language. Whenever an observation can be tied to source text, the report includes a short excerpt so the reader can reopen the conversation and check what was actually written. See how the DryText report works and the supported WhatsApp chat analyzer.

Readers can evaluate an observation in three steps: check whether the excerpt matches the source, read the surrounding messages, and compare the pattern with the question they asked. A visible count, such as who initiated more threads in the selected period, is different from an explanation of why that happened. If the excerpt does not support the wording, or the selected messages omit important context, treat the conclusion as incomplete and rely on your own knowledge of the conversation.

That sequence matters: a count or pattern is an observation, while an explanation is an interpretation. Confidence labels communicate when an interpretation is tentative. They do not turn an uncertain inference into proof. The original conversation and the people who know its context remain the best place to check whether a report makes sense.

What DryText does not do

• It does not log into WhatsApp or iMessage, retrieve deleted messages, or access a phone account.
• It does not analyze photos, videos, voice notes, calls, body language, or events that were never written in the export.
• It cannot prove attraction, dishonesty, commitment, personality, or someone’s intent from message length or reply speed.
• It is not a therapist, diagnostic service, relationship authority, or substitute for professional support.
• It should not be used to monitor another person secretly, pressure them, or make a high-stakes decision without other evidence.

Privacy, retention, and deletion

Only upload a conversation you are entitled to use. Before uploading, review the participant names and date range, remove unrelated information where practical, and consider whether everyone in a group chat would reasonably expect the conversation to be analyzed. Uploaded text is processed to create the report and may be sent to the configured AI provider.

Transcripts without a report are deleted within 48 hours. Transcripts associated with a report are deleted within seven days of the latest report. Users can delete chats and reports sooner from their account; account and payment records may follow separate retention requirements. The safety and data-retention guide explains the product controls, and the Privacy Policy describes personal-data handling in more detail.

Product background and contact

DryText launched in 2026 as an independent web app for reviewing supported chat exports. It is not affiliated with WhatsApp, Meta, Apple, or iMessage. The product is maintained around a narrow goal: make patterns easier to inspect while keeping uncertainty, source excerpts, and privacy limitations visible.

You can contact the team without signing in at support@drytext.app. For account or report questions, the support page includes self-service paths. We aim to respond to support requests within two business days.