How is the script called?

[hdprice]

I've read through the docs, and I can't discover how the script is called. Do I need to create an automation to call it? If so, can you describe how to do that and what parameters need to be passed?

[michi1996]

I have the same problem. The option 3 script will not geting called from open ai integration.

[nikito]

I've seen that it is called for most of my commands, but some are going to the Hass script instead. May be good to offer feedback to the The HA team on things not working with the Hass script to help them make it better. Not sure what else can be done from this side, it may have to do with the intent tied to the script handling it before the llm, or just the script itself seeming like a better match to the llm. Could try experimenting with the script description to see if it can coax the ai to prefer it over others.

[Jack77777777]

Has someone got this fixed. The Option 3 script is working intermittent for me.

[hdprice]

Has anybody had any luck figuring this out?

[nikito]

Can try playing with the script description this see if it can be given more preference, but seems the hass script is more favored by the llms 🫤

[Jack77777777]

Tried Today to call Option 3 script several times with no luck. Is there a way to disable the Hass script?

[Jack77777777]

Can try playing with the script description this see if it can be given more preference, but seems the hass script is more favored by the llms 🫤<

Thank you in advance. I can test if needed.
Just for reference: my used language is german

[thrunkmo]

I added this to my Claude AI prompt:

The "script.llm_script_for_music_assistant_voice_requests" script is used to play music based on a voice request. The tool takes the following arguments: media_type, artist, album, media_id, radio_mode, area. media_id and media_type are always required and must always be supplied as arguments to this tool. An area or Music Assistant media player can optionally be provided in the voice request as well. Use the parameters as described in the description of each parameter. Use this tool whenever the user requests to play music.

[Jack77777777]

I added this to my Claude AI prompt:

The "script.llm_script_for_music_assistant_voice_requests" script is used to play music based on a voice request. The tool takes the following arguments: media_type, artist, album, media_id, radio_mode, area. media_id and media_type are always required and must always be supplied as arguments to this tool. An area or Music Assistant media player can optionally be provided in the voice request as well. Use the parameters as described in the description of each parameter. Use this tool whenever the user requests to play music.

Is the script called each time with this description for you?

[thrunkmo]

I added this to my Claude AI prompt:
The "script.llm_script_for_music_assistant_voice_requests" script is used to play music based on a voice request. The tool takes the following arguments: media_type, artist, album, media_id, radio_mode, area. media_id and media_type are always required and must always be supplied as arguments to this tool. An area or Music Assistant media player can optionally be provided in the voice request as well. Use the parameters as described in the description of each parameter. Use this tool whenever the user requests to play music.

Is the script called each time with this description for you?

Yes, it seems to consistently call it. I just tested it with "Play AC/DC in the office" and "Play AC/DC on the office speaker" and it immediately called it both times. I'll keep an eye on how consistent it is and update here in the future if I remember to

[Jack77777777]

@thrunkmo WOW! It seems to work.
Thank you a lot!

[thrunkmo]

@thrunkmo WOW! It seems to work. Thank you a lot!

Is yours still working? Need to fiddle with mine. My Voice PE says it's calling the script to play music on my specific speaker, but nothing ever plays now. Just happened in the last week. Will do some troubleshooting and post back here if I get it fixed

[SteveDinn]

Is your music source Spotify or YouTube music? There were some changes put into music assistant recently that fixed a change that Spotify made. Try updating.

[thrunkmo]

Is your music source Spotify or YouTube music? There were some changes put into music assistant recently that fixed a change that Spotify made. Try updating.

EDIT: FIXED (PEBCAK issue). I changed my queue settings in MA and my Voice PE was queuing thigns up but not playing them. And I needed to re-authenticate Spotify. D'oh. Origional post below:

Appreciate the comment, but yes, I've used the DNS fix and now the 2.5.8 MA update which fixes Spotify for me. I don't think it's that

I even get a confirmation from Voice PE: "Colter Wall's music is now playing in the Office. Enjoy." Just no music. Hmm. Looking at my Trace now:

AH! Found an issue from MA logs and will track this down. Doesn't seem to be an LLM issue:

2025-08-11 14:07:01.914 INFO (MainThread) [music_assistant] Loaded plugin provider Home Assistant
2025-08-11 14:25:49.963 INFO (MainThread) [music_assistant.player_queues] Fetching radio tracks for queue Office based on: Colter Wall
2025-08-11 14:25:50.070 INFO (MainThread) [music_assistant.spotify] Attempt 1/5 failed: Token expired
2025-08-11 14:25:50.070 INFO (MainThread) [music_assistant.spotify] Retrying in 0.05 seconds...
2025-08-11 15:10:30.209 INFO (MainThread) [music_assistant.player_queues] Fetching radio tracks for queue Office based on: Colter Wall
2025-08-11 15:13:11.106 INFO (MainThread) [music_assistant.players] Player registered: ma_xqrh2cewvi/This Device

[Coxeroni]

I added this to my Claude AI prompt:

The "script.llm_script_for_music_assistant_voice_requests" script is used to play music based on a voice request. The tool takes the following arguments: media_type, artist, album, media_id, radio_mode, area. media_id and media_type are always required and must always be supplied as arguments to this tool. An area or Music Assistant media player can optionally be provided in the voice request as well. Use the parameters as described in the description of each parameter. Use this tool whenever the user requests to play music.

Where do I put that?
In the general Anthropic/Claude settings in the intergation settings?