Conversational AI - Beta Release

ElevenLabs Conversational AI is now available in beta, allowing you to design and engage with personalized AI voice agents.  Pick a voice, select an LLM (with the option to integrate your own in the near future), craft a system prompt, and initiate your conversation. This allows you to develop interactive support agents, video game characters, math tutors, and more using ElevenLabs voices.  

Since the initial alpha release, ElevenLabs has added some new features to Conversational AI.

• Knowledge Base for AI Agents - Files can be uploaded or text extracted from URLs to comprise the AI Agent's knowledge base. Non-enterprise users are limited to 5 files or links and up to 20MB total. Need more? Contact our sales team to discuss an enterprise plan for higher limits.
• Evaluation Criteria - set target criteria for the call (e.g. "Has the agent answered all the questions?")
• Data extraction - extract data about the conversation, e.g. type of request "information_seeking", "feature request", etc.
  
• SDK guides as well as general information about Conversational AI is now available in the ElevenLabs documentation at https://elevenlabs.io/docs/conversational-ai/overview
• Pricing: 1000 credits per minute + LLM costs (note:  LLM usage is not currently in effect but will be in the near future, at the moment only the 1000 credits per minute charge is in effect.)
  

As Conversational AI is in its early stage, you may encounter bugs or notice some features are not yet available.  You can report any in-call bugs directly through the platform or voice your thoughts and suggestions in the ElevenLabs Discord (Discord invite link is discord.gg/elevenlabs).