DocVoice — Talk to Your Documents

DocVoice is an AI document assistant that lets you have a spoken, real-time conversation with your PDFs instead of reading or typing queries. You upload a PDF, it is processed and indexed, and then you talk to it and hear answers back. It is built for anyone who needs to extract information from documents conversationally, including businesses that want visitors to ask questions about an annual report or manual directly on their website. Concrete features include support for PDFs up to 50 MB, API responses generated in under one second, and full-duplex streaming via Gemini Live with natural interruption handling so you can talk over the assistant. Under the hood it uses Pinecone vector indexing for semantic search and Gemini 2.5 Flash for responses, with per-account data isolation and 'origin-locked' keys that restrict embeds to specific domains. A single-line embed code adds the voice interface to any website, and both voice and text input are supported. DocVoice is free to start with no credit card required; full pricing is not published on the page.

People use DocVoice to talk to their PDFs by voice and get spoken answers, or to embed a voice Q&A assistant on a website.