I have Piper set up and configured in HA 2026.4.1. If I go into Media from the nav bar, Text-to-speech is one of my available sources. If I click it, I can pick between Google and Piper. Picking Piper I then get a box where I can type some text in that I want it to say, and I can pick the language and voice, and down in the bottom right, I can pick the source that will play the text. All pretty standard stuff.
After typing in some text, picking an ESPHome (older build; maybe 2025 something) media speaker source (which btw seems to reset everything), I hit Say and it proceeded to play the text… and loop endlessly repeating the same thing with about a 15-30s delay between them. I noticed that the media player was stuck in “playing” status, and I had to go to developer actions and issue a media player stop command to get it to stop looping.
Has anyone else noticed this? I didn’t used to do this; I know because I use this every once in a while to announce things to my kids in other rooms. It used to just say the text once and stop.
I also noticed that the language and voice options do not match the language/voice set in Settings - Voice assistants - Home Assistant (my only voice assistant configured). It would be nice if they matched, because I’m very much guessing that if they don’t match, it will cause Piper to download the required voice files and I’d rather it not do that. Also, having to pick the right language/voice every time that I like is frustrating, especially when the forms reset whenever you pick a different source.
I’m wondering if the audio looping might be caused by my media player being built on an older ESPHome release, or if it’s Home Assistant itself.