Mother’s Voice AI Chatbot
Budget: ₹600 – ₹1,500 INR
I have several hours of phone recordings of my late mother and I want to turn them into a living, talking keepsake. The goal is a private-use voice chatbot that sounds exactly like her, can hold free-flowing conversations with me, and, when prompted, read messages or short stories in her distinctive tone.
Here’s what matters most to me:
• Voice fidelity must be uncanny—when I close my eyes it should feel as though she is really there.
• Conversations need to flow naturally, without awkward pauses or robotic phrasing.
• The system should run on-device or behind a simple web/mobile interface so I can speak or type and hear her answer in real time.
I will supply the raw recordings and any additional text data you need for fine-tuning. You can choose the cloning and language-model stack (e.g., ElevenLabs, Azure Neural TTS, OpenAI, custom Tacotron + GPT) as long as the final result meets the realism and latency goals.
Deliverables:
1. Trained voice model and documented training pipeline.
2. Conversational agent capable of open-domain chat plus a “read this” mode for stories or messages.
3. Simple UI (CLI, web, or mobile) with install/run instructions.
4. Short demo video proving natural interaction.
If you have prior examples of ultra-high-fidelity voice cloning and dialogue work, please share them—hearing is believing.
Here’s what matters most to me:
• Voice fidelity must be uncanny—when I close my eyes it should feel as though she is really there.
• Conversations need to flow naturally, without awkward pauses or robotic phrasing.
• The system should run on-device or behind a simple web/mobile interface so I can speak or type and hear her answer in real time.
I will supply the raw recordings and any additional text data you need for fine-tuning. You can choose the cloning and language-model stack (e.g., ElevenLabs, Azure Neural TTS, OpenAI, custom Tacotron + GPT) as long as the final result meets the realism and latency goals.
Deliverables:
1. Trained voice model and documented training pipeline.
2. Conversational agent capable of open-domain chat plus a “read this” mode for stories or messages.
3. Simple UI (CLI, web, or mobile) with install/run instructions.
4. Short demo video proving natural interaction.
If you have prior examples of ultra-high-fidelity voice cloning and dialogue work, please share them—hearing is believing.