56 messages, 4 voice notes, 48 seconds: the briefing that reads my family's WhatsApp while I sleep

Description and use

Amar Khobor turns overnight family WhatsApp voice notes into a 48‑second warm audio briefing with structured show notes, saving users from manually listening to dozens of messages.

Problem

Family WhatsApp groups accumulate many overnight voice notes in multiple languages, making it hard to catch up quickly.

Outcome

The tool produced a 48‑second warm audio briefing and structured show notes that accurately captured the content of four voice notes, saving the user time and reducing transcription errors.

Workflow

• Export WhatsApp group chat
• Apply energy‑based VAD to filter silence
• Transcribe audio with Sarvam Saaras
• Generate summary JSON with Gemma 4
• Extract spoken script and show notes
• Narrate script with ElevenLabs
• Deliver audio briefing and show notes

Evidence

Amar Khobor turns a chaotic overnight WhatsApp group — dozens of voice notes in Bengali, Hindi, English, Benglish, and Hinglish — into one warm, two‑minute audio briefing you can play while making morning tea.

Source

Whisper wrote "trouble at the market" (হাটে সমসা). My father said "heart problem" (হার্টের সমস্যা).

Source

The fix was a tolerant backward extractor: find the last occurrence of "spoken_script", backtrack to the { that opens that object, and slice to the closing }.

Source