Telegram Voice Note and Audio Transcription using Google Gemini - n8n Workflow

Users who frequently receive voice notes and need instant, searchable transcripts.
Developers building conversational AI tools leveraging Telegram and LLMs.
n8n users looking for advanced examples of integrating multimodal AI models like Gemini via custom HTTP requests.
Automation specialists seeking effective n8n templates for handling media files.

Handling voice notes can be time-consuming, especially when needing to extract key information. This robust n8n workflow solves this by creating a seamless bridge between your Telegram bot and Google Gemini’s powerful transcription capabilities. The automation begins with a dedicated n8n trigger monitoring incoming Telegram messages. If an audio file is detected, the workflow intelligently routes the data, uploads the file via a sequence of HTTP request n8n nodes, sends it to Gemini for transcription, and finally posts the readable text back to the original chat. This is a crucial example of an aiautomationworkflow utilizing custom scripting within the n8n environment to achieve complex integrations.

This automation starts with the Telegram Trigger1 n8n trigger, which listens for any new messages sent to your bot.

1. Incoming Message Handling: The data flows to a Switch n8n node. This corelogicflow_control mechanism determines if the incoming message contains an audio file (to be transcribed) or simple text (to be ignored or routed elsewhere).


2. File Preparation: If an audio file is present, the workflow proceeds to initialize the file upload using the first HTTP Request n8n node (initialize upload session). This is typically necessary to prepare a temporary upload location for the AI service.


3. File Upload: A subsequent Merge node combines necessary file metadata, leading to the Upload file HTTP Request n8n node, which pushes the actual audio data to the designated temporary cloud location.


4. Gemini Transcription: Once uploaded, the Ask Gemini to transcribe HTTP Request n8n node is executed. This critical step sends the file URI, along with the transcription prompt, to the Google Gemini API.


5. Output and Reply: The transcribed text returned by Gemini is then immediately formatted and sent back to the user using the Reply in Telegram n8n node, completing this powerful n8n workflow.

To install and run this n8n workflow successfully, follow these steps:

1. Import the n8n Workflow: Copy the provided JSON data and paste it into your n8n instance via the 'New' menu -> 'Import from JSON'.


2. Set Up Telegram Credentials: Configure the Telegram Trigger1 n8n trigger and the Reply in Telegram n8n node with your Telegram Bot credentials (API Key and Chat ID, if applicable).


3. Configure Gemini API: The HTTP Request nodes (initialize upload session, Upload file, and Ask Gemini to transcribe) require configuration specific to the Gemini service. You must ensure your API key is correctly authenticated, either globally or within the node's settings, and that the URLs are correctly pointed toward the Gemini file upload and processing endpoints.


4. Test the n8n Trigger: Save the n8n workflow and activate it. Send a voice note to your Telegram bot to test the transcription process.

Telegram Trigger1 (n8n trigger): The starting point of the n8n workflow. It is configured to listen for incoming messages and files sent to the associated Telegram bot.
Switch (n8n node): A corelogicflow_control node that evaluates the incoming message structure to determine if it's an audio file or simple text, routing the data accordingly.
initialize upload session (HTTP Request n8n node): Sends an initial request to the Gemini-compatible service endpoint to secure a temporary session ID and URL for the incoming audio file.
Upload file (HTTP Request n8n node): Responsible for the actual transfer of the binary audio file data retrieved from Telegram to the temporary upload URL established in the previous step.
Ask Gemini to transcribe (HTTP Request n8n node): The core AI processing node. It sends a request to the Gemini API, referencing the uploaded file's URI and instructing the model to generate a full transcription.
Reply in Telegram (n8n node): The final action. It uses the chat context from the initial n8n trigger to send the resulting transcribed text back to the user via Telegram.