🎙️ VoiceScribe AI: Telegram Audio Message Auto Transcription with OpenAI Whisper

Automatically transcribe Telegram voice messages and store them as structured logs in Google Sheets, while backing up the audio in Google Drive.

🧑‍💼 Who’s it for

• Journalists, content creators, or busy professionals who often record voice memos or short interviews on the go.
• Anyone who wants to turn voice recordings into searchable, structured notes.

⚙️ How it works / What it does

1. User sends a voice message to a Telegram bot.
2. n8n checks if the message is an audio voice note.
3. If valid, it downloads the audio file and:
   • Transcribes it using OpenAI Whisper (or your LLM of choice).
   • Uploads the original audio to Google Drive for safekeeping.
4. The transcript and audio metadata are merged.
5. The workflow:
   • Logs the data into a Google Sheet.
   • Sends a formatted confirmation message to the user via Telegram.

If the input is not audio, the bot politely informs the user that only voice messages are accepted.

✅ Features

• Accepts only Telegram voice messages.
• Transcribes via OpenAI Whisper.
• Logs DateTime, Duration, Transcript, and Audio URL to Google Sheets.
• Sends user feedback message via Telegram with download + transcript link.

🚀 How to set up

Prerequisites

• Telegram Bot connected to n8n (via Telegram Trigger)
• Google Drive & Google Sheets credentials configured
• OpenAI or Whisper API credentials (for transcription)

Steps

1. Telegram Trigger
   Start the flow when a new message is sent to your bot.
2. Check Message Type
   Use a conditional node to confirm it's a voice message.
3. Download Voice Message
   Download the .oga file from Telegram.
4. Transcribe Audio
   Send the binary audio to OpenAI Whisper or your transcription service.
5. Upload to Google Drive
   Backup the original audio file.
6. Merge Outputs
   Combine transcription with Drive metadata.
7. Transform to Row Format
   Prepare structured JSON for Google Sheets.
8. Append to Google Sheet
   Store the transcript log (DateTime, Duration, Transcript, AudioURL).
9. Send Confirmation to User
   Inform the user via Telegram with their transcript and download link.
10. Unsupported Message Handler
    Reply to users who send non-audio messages.

📄 Example Output in Google Sheet

🧠 How to customize the workflow

• Swap Whisper with Deepgram, AssemblyAI, or other providers.
• Add speaker name detection or prompt-based tagging via GPT.
• Route transcripts into Notion, Airtable, or CRM systems.
• Add multi-language support or summarization steps.

📦 Requirements

Created with ❤️ using n8n