This n8n workflow automates the process of transcribing a voice message received via Telegram, summarizing the transcription using an AI model, and then saving both the transcription and the summary to separate Google Docs in Google Drive.
Here's a breakdown of the workflow:
- Telegram Trigger: The workflow is initiated when a new message is received on Telegram.
- Telegram Node: It retrieves the voice message file from Telegram using the
file_id
from the incoming message.
- OpenAI Node: The retrieved voice file is then sent to OpenAI for transcription.
- Google Drive (for Transcription): The transcribed text is saved as a new Google Doc with the current time as its name in a Google Drive folder named "N8N Transcribes".
- AI Agent: The transcribed text is sent to an AI Agent (connected to a DeepSeek Chat Model) to generate a plain text summary. The prompt for the summary is set to "Generate A summary description from this Telegram message. Don't use ** or anything that emulates rich text. The description should be in plain text Expand on the title so that the description can be used in a report to my supervisor. Only provide one summary. Do not label it Summary*, I already know it's a summary. Thanks.:{{ $json.text }}".
- Google Drive (for Summary): The generated summary is then saved as a new Google Doc with the current time as its name in a Google Drive folder named "N8N Summaries".
In essence, this workflow allows users to send a voice message to a Telegram bot, and automatically get both a transcription and a summary of that message saved as separate documents in their Google Drive.