Telegram Audio Bot with ElevenLabs + Make.com (2025) · published 24 August 2025
Play at 0:47 Watch on YouTube Telegram
How do I transcribe a Telegram voice note with ElevenLabs in Make.com?
I download the Telegram voice note and POST it to the ElevenLabs speech-to-text URL. The file name has to end in .oga or the call errors.
This continues the Telegram bot from the earlier video. Watch updates, pick the webhook, save. Telegram Download a file, same connection. The file ID is the message attachment file ID, not a photo ID. Save.
The next module is HTTP, Make a request, with API-key credentials. If I do not have a key yet, I add one: ElevenLabs profile, API keys, create a key with the restrictions removed so I can reuse it, copy it once, and paste it into Make. The URL is the speech-to-text URL from the ElevenLabs docs, the one on screen. I put the method on GET in the video. That endpoint only accepts POST, so GET does not upload the file. The header is the API-key header again, with the same key. Body type is multipart/form-data. Field type is file, key is file, and I map the Telegram download. On the file name I add .oga, because that is what Telegram outputs. I turn on parse response and save.