Hello, I am looking for a developer to build a live interactive AI television host using HeyGen LiveAvatar. This is NOT a pre-recorded avatar/video project. The AI presenter will participate in an actual live television broadcast and communicate with viewers calling the show by telephone. The TV show and broadcasting infrastructure already exist. vMix is already installed and currently used for live broadcasting. The required workflow is approximately: Viewer phone call → Estonian speech recognition → AI/LLM → natural Estonian voice → HeyGen LiveAvatar → vMix → live television The first version should use one avatar. Main requirements: • HeyGen LiveAvatar API / SDK integration • Incoming telephone calls (Twilio, Vapi or another reliable and cost-effective solution) • High-quality Estonian speech recognition • OpenAI or another suitable LLM for conversation • Conversation context/memory • Natural Estonian TTS (ElevenLabs or another suitable solution) • Real-time avatar lip-sync and movement • Integration of the avatar video/audio output with our existing vMix broadcasting system • Approximately 3–5 seconds end-to-end response latency is acceptable • High reliability for live television • Low ongoing operating/API costs are important IMPORTANT – BEFORE WE START Please tell me clearly in your first reply whether you charge for the initial discussion/consultation. I am not looking for paid consultation at this stage. I need an initial free discussion to explain and refine the idea, determine whether you are technically able to build it, and agree on the proposed architecture and approximate scope. If we both agree that the project is technically feasible and decide to proceed with development, we will then agree on the development price/payment terms before any paid work begins. I am not asking you to perform development work for free. I simply do not want to enter a paid consulting arrangement from the first conversation before we have established whether you are the right developer for the project. If you are interested, please answer:

Have you worked specifically with HeyGen LiveAvatar API/SDK?

Can you show me a working LiveAvatar or real-time AI avatar example?

Have you built real-time AI voice/telephone agents?

How would you connect the LiveAvatar video/audio output to vMix?

Can you support Estonian STT and Estonian TTS?

What end-to-end latency would you expect?

Approximately how long would the MVP take to build?

What approximate development budget would you expect?

What would you estimate the ongoing operating/API cost per hour of live broadcasting to be? The objective is to first build a reliable MVP with one avatar, test it in the real television environment, and expand it later if successful. If needed, contact me via saade@saade.ee e-mail. Best regards Madis Sutt