I put together a small project that lets you talk to an AI Santa Claus at https://talktosanta.io. The voice is generated using ElevenLabs (based on a trained voice actor), and there’s no login required. For each conversation started, ElevenLabs will donate $2 to charity (capped at $11,000).
The code is open-source: https://github.com/elevenlabs/elevenlabs-examples/tree/main/...
This is cool! Do you know to add the Santa voice to a newly created Agent? It's not showing up on the ElevenLabs UI. Can't get the proper santa AGENT_ID working.
The 11labs agent product is promising but very buggy, I was glad to see some of those issues in your implementation as well. The agent often drops out and comes back randomly after long stretches of silence. They also seem to be running some kind of half-functional censor on the llm output when used with Gemini 1.5. It also bills 500 credits for any time < 1 minute, not the aggregate use.
The code is open-source: https://github.com/elevenlabs/elevenlabs-examples/tree/main/...
This is cool! Do you know to add the Santa voice to a newly created Agent? It's not showing up on the ElevenLabs UI. Can't get the proper santa AGENT_ID working.
Thanks! We created a custom professional voice clone for this purpose using a voice actor. Let me check with Santa to see if we have permission to make his voice public :)
Fun! The 11labs agent product is promising but very buggy, I was glad to see some of those issues in your implementation as well. The agent often drops out and comes back randomly after long stretches of silence. They also seem to be running some kind of half-functional censor on the llm output when used with Gemini 1.5. It also bills 500 credits for any time < 1 minute, not the aggregate use.
Thanks for the feedback - What do you mean by the censor? As in safety moderation? For this specific agent I had to prompt it very carefully to makes sure people wouldn't ask Santa for anything dangerous :)
Billing component is noted, round-up doesn't seem fair. Let me check this.
Fun! The 11labs agent product is promising but very buggy, I was glad to see some of those issues in your implementation as well. The agent often drops out and comes back randomly after long stretches of silence. They also seem to be running some kind of half-functional censor on the llm output when used with Gemini 1.5. It also bills 500 credits for any time < 1 minute, not the aggregate use.
Thanks for the feedback - What do you mean by the censor? As in safety moderation? For this specific agent I had to prompt it very carefully to makes sure people wouldn't ask Santa for anything dangerous :) Billing component is noted, round-up doesn't seem fair. Let me check this.
Thanks. Censor – it's hard to say what's happening because it's not even carried through in the transcription, but the agent will say one word at the start of a natural sentence and then stop talking. Seems to only happen with Gemini.
Thanks for the feedback - What do you mean by the censor? As in safety moderation? For this specific agent I had to prompt it very carefully to makes sure people wouldn't ask Santa for anything dangerous :) Billing component is noted, round-up doesn't seem fair. Let me check this.
Thanks. Censor – it's hard to say what's happening because it's not even carried through in the transcription, but the agent will say one word at the start of a natural sentence and then stop talking. Seems to only happen with Gemini.
Very strange - let me look into this. Thanks for flagging.
This is cool! Do you know to add the Santa voice to a newly created Agent? It's not showing up on the ElevenLabs UI. Can't get the proper santa AGENT_ID working.
Thanks! We created a custom professional voice clone for this purpose using a voice actor. Let me check with Santa to see if we have permission to make his voice public :)