Summarise WhatsApp Chats with AI (Step-by-Step) | ThreadRecap
Summarising a long WhatsApp thread by hand, scrolling, copying, pasting, and prompting, is the slow path. The clean workflow is: export the chat, include voice notes if they matter, upload the `.zip`, and let a tool that understands WhatsApp do the structural work.
This page walks through the full process end-to-end, including the parts most guides skip, voice note transcription, group chat filtering, credit math, and the privacy details that matter when you upload someone else's words to an AI tool.
The fastest workflow
Export the chat from WhatsApp as a `.zip` (with media if you want voice notes transcribed).
Upload it to a tool that natively understands the WhatsApp export format.
Pick the output you want: Summary, Decisions, Action Items, Open Questions, Meeting Recap.
Get a structured recap in minutes. Ask follow-up questions inline, or export decisions and action items to Notion, Trello, or Google Calendar with one click.
If you already have a WhatsApp export `.zip`, you are one upload away from a clean recap.
What you will get
A typical recap returns the following sections, with the exact set depending on the goal you select:
Summary: narrative overview of the conversation.
Decisions made: what was agreed, who decided, and the supporting context.
Action items: tasks with owners, deadlines (where mentioned), and current state.
Open questions: unresolved topics that need a follow-up.
Notable quotes: direct quotes worth preserving, with timestamps.
Suggested follow-ups: concrete next steps the analysis recommends.
After the recap is generated, you can ask follow-up questions through the in-app AI chat to clarify details or pull exact quotes. Decisions and action items can be sent directly to Notion, Trello, or Google Calendar with one click.
Step 1: Export your WhatsApp chat (get the `.zip`)
Choose Include media if you want voice notes transcribed. Without media if text-only is enough.
Save or share the exported `.zip`.
iPhone
Open the chat.
Tap the contact or group name at the top.
Scroll down to Export Chat.
Choose Attach Media for voice notes, Without Media for text-only.
Save or share the `.zip`.
Should you choose "Include media"?
Choose it if you have substantive content in voice notes, decisions, rationale, action items, anything spoken rather than typed. ThreadRecap transcribes the voice files and merges them into the timeline so the recap reflects the full conversation.
Choose "Without media" if the chat is mostly text and you want the smallest file. Without-media exports also have a higher message cap (40,000 vs 10,000 with media), which matters for long historical chats.
Each goal returns the same structure on every run, so weekly recaps for the same project look consistent and you can compare them over time. When the chat is a recurring work sync, the Meeting Recap goal lands directly as meeting minutes from a WhatsApp chat.
Custom prompts
If the goal templates do not match your use case, you can write a custom prompt (+3 credits). The custom prompt sits on top of the same parsed conversation, so participant filtering, date ranges, and voice transcription still apply.
Step 3: Voice notes to text (the part most people skip)
WhatsApp conversations often hide the substantive content in voice notes. A 3-minute voice note typically covers more ground than 50 text messages, especially in workflows where the people doing the substantive talking prefer to dictate.
ThreadRecap handles voice transcription as part of the same upload:
Supported formats: `.opus` (the WhatsApp default) and `.m4a` (older iOS exports).
Transcription engine: OpenAI Whisper.
Accuracy: approximately 95% on clear audio. Background noise, overlapping speech, and very low-volume recordings reduce that figure to 80–90%.
Timeline merge: each transcribed clip is inserted into the conversation at its original timestamp, attributed to the original sender, so the analysis treats audio identically to typed messages.
Without timeline merge, even a tool that "transcribes voice notes" will miss the substantive content because the analysis layer never sees the audio in conversation context. ThreadRecap's pipeline does the merge automatically.
Practical tip: if your chat relies on voice notes, export with media. Otherwise you are summarising only half the conversation.
Step 4: Understand cost (credits) before you run analysis
ThreadRecap uses credits and charges per usage:
1 credit per 1,000 messages (rounded up).
1 credit per 10 minutes of audio (rounded up).
Plus modifiers:
Paid analysis goals: +2 credits.
Custom prompt: +3 credits.
Group chat analysis: +2 credits.
Translation to English: +1 credit.
New users get 5 free credits on sign-up, which is enough to run a complete recap on a typical short or medium chat end-to-end before deciding whether to top up.
Credit math examples
Scenario
Credits
800-message one-on-one chat, no voice notes, Summary goal
1
3,500-message group chat, 20 min voice notes, Meeting Recap
4 (messages) + 2 (audio) + 2 (group) = 8
10,000-message chat, 60 min voice notes, Action Items + Decisions on same upload
10 + 6 + 10 (second goal reuses parse) = 26
5,000-message group chat, custom prompt, no audio
5 + 3 + 2 = 10
Multiple goals on the same upload reuse the parse and transcription work; only the per-message credit applies again.
Step 5: Group chats (keep it focused)
Group chats explode token usage and signal-to-noise ratio. ThreadRecap's approach is to focus analysis on the participants whose contributions actually matter:
Participant filtering: exclude reactions, jokes, side conversations.
Date range filtering: focus on the period that matters for your task.
Goal selection: Meeting Recap or Action Items beats generic Summary for work chats.
In a 12-person group chat, the three or four people doing 80% of the substantive talking are usually the only ones whose messages need to enter the analysis. Filtering down sharpens the recap and reduces credit cost.
Privacy: what happens to your data
Specific privacy claims:
Local parsing: `.zip` files are unzipped and parsed in your browser. Photos, videos, and documents never leave your device.
Selective upload: only chat text and voice note audio are sent to ThreadRecap's servers, the rest of the archive stays on your device.
Encrypted account storage: chat text, voice note audio, and processed recaps are stored encrypted in your account so you can return to the AI chat and replay clips. You control deletion through the dashboard at any time.
No data sale: ThreadRecap states it does not sell user data.
No model training on user content: ThreadRecap states it does not train AI models on your chat content.
For sensitive conversations (legal, medical, HR, family disputes), review the privacy policy before uploading. The retention specifics and deletion behaviour are detailed there.
Common problems and fixes
"I cannot export this chat"
WhatsApp's Advanced Chat Privacy feature can block export on a per-chat basis. The setting lives inside the individual chat, not at the device or account level. If export is blocked, someone in the chat enabled the restriction; you cannot bypass it from your side.
"My export is missing voice notes"
You exported without media. Re-export and choose Include media (Android) or Attach Media (iPhone).
"The `.zip` is huge"
Large media exports can be hundreds of megabytes or more. If you only need a text summary, export without media for a smaller file. If you need audio transcription, keep media but consider exporting a smaller date range, especially if the conversation spans years.
"The recap missed an important decision"
If the decision was made in a voice note and you exported without media, that decision is not in the chat log at all, the export shows "audio omitted". Re-export with media and re-run the analysis.
If the decision was in text but the recap missed it, the issue is usually filter-related: a date range that excluded the decision, or a participant filter that excluded the speaker. Adjust filters and re-run.
"My chat is too long for ChatGPT but I want a recap"
That is exactly the case ThreadRecap is built for. See summarise long WhatsApp chats for the full multi-pass workflow at 5k, 10k, and 50k message scale.
Quick reference
Can AI summarise an entire WhatsApp conversation?
Yes. The reliable approach is to export the chat as a `.zip` and run analysis on the export rather than scrolling and copying sections manually. A purpose-built tool handles the full chat in one pass and preserves coherence across thousands of messages.
Do I need to include media?
Only if voice notes carry meaningful content. ThreadRecap supports `.opus` and `.m4a` audio and merges transcripts into the conversation timeline.
What file types are inside the export?
`_chat.txt` (the full message history) plus media files (`.jpg`, `.mp4`, `.opus`, `.m4a`, etc.) when media is included.
Is it safe to upload a WhatsApp export?
ThreadRecap parses `.zip` files locally in the browser, never uploads photos, videos, or documents, and stores chat text, voice note audio, and processed recaps encrypted in your account where you control deletion through the dashboard. Review the privacy policy before uploading sensitive conversations.
What languages are supported?
The UI is available in multiple languages and ThreadRecap can return outputs in the conversation's original language or in English. Voice transcription works across the 50+ languages Whisper handles well.
Summarize WhatsApp chat conversations instantly with AI—extract summaries, decisions, action items, and open questions from exported .zip files in minutes.
Nov 6, 202511 min read
Ready to analyze your WhatsApp chat?
Upload your export and get summaries, insights, and voice note transcriptions in minutes.
