tl;dr as of December 17, 2025, Microsoft has blocked our method of being able to get free text-to-speech
--
Hey folks, I'm working on trying to figure out the best way to get the word across — but for now, here's the state of things:
The long and short of it is that as of now, in order to be able to emulate the Edge web browser and request (free) text-to-speech from Microsoft servers, part of the emulation includes adding special headers to the websocket request.
But browser environments do not allow for custom websocket headers.
I have been spinning my wheels for the past couple of days trying out work-arounds and seeing if I could manage to get something to work — but I don't see a way around this restriction.
For now, you can still get edge-tts through formats like the original Python package rany2/edge-tts and I managed to make it work for the free edge-tts mobile app on Android and iOS
--
Hey folks, I'm working on trying to figure out the best way to get the word across — but for now, here's the state of things:
The long and short of it is that as of now, in order to be able to emulate the Edge web browser and request (free) text-to-speech from Microsoft servers, part of the emulation includes adding special headers to the websocket request.
But browser environments do not allow for custom websocket headers.
I have been spinning my wheels for the past couple of days trying out work-arounds and seeing if I could manage to get something to work — but I don't see a way around this restriction.
This has taken out — overnight — my edge-tts browser extension, Obsidian plugin, and web interface https://tts.best/
For now, you can still get edge-tts through formats like the original Python package rany2/edge-tts and I managed to make it work for the free edge-tts mobile app on Android and iOS
To anyone interested, the line in the Python code that is basically "missing" from my attempts to fix the various packages and services, it's here: https://github.com/rany2/edge-tts/blob/master/src/edge_tts/communicate.py#L441