# VoiceClone AI — Full Description > Free in-browser AI voice cloning demo by M Asif Rahman. ## What it is VoiceClone AI is a single-page web application that demonstrates real-time voice cloning. A visitor records a short English voice sample directly in the browser, picks one curated English script from a fixed catalog, and receives a synthesized audio clip of that script spoken in their own cloned voice. The product is a free technology demo, not a commercial service. It is intended for personal experimentation, learning, and showcasing what modern voice synthesis can do. ## How it works 1. The visitor grants microphone access and records a short voice sample (a few seconds of clear English speech). 2. The visitor selects one script from a curated catalog of pre-approved English texts (e.g. famous movie quotes, narration excerpts, character lines). 3. The recording is sent to a backend voice synthesis pipeline together with the chosen script identifier. 4. A cloned audio clip is generated and streamed back into the browser, where it can be played and downloaded. ## Constraints (deliberate) - **English only.** Both the voice sample and the script must be in English. Requests in other languages are rejected server-side. - **No free-text input.** Visitors cannot type their own text to synthesize. Only the curated catalog of scripts is accepted; arbitrary text submissions are blocked at the API layer. - **Maximum input length: 400 characters** per script. - **Voice samples are deleted immediately** after the synthesis request completes. Only the generated audio output is retained. - **Rate-limited.** Each visitor is limited to a small number of generations per time window to keep the demo fair and prevent abuse. - **Demo only.** No accounts, no payments, no SLA. The service may change or go offline at any time. ## Pages - `/` — Home: record, select a script, generate, and play back the cloned audio. ## Endpoints - `/sitemap.xml` — XML sitemap of public URLs. - `/llms.txt` — Short llmstxt.org-format index for AI crawlers. - `/llms-full.txt` — This document. - `/robots.txt` — Crawl directives, including explicit allow rules for major AI agents. - `/.well-known/api-catalog` — Machine-readable list of public, AI-relevant endpoints. ## Audience Developers, designers, AI enthusiasts, and curious visitors who want to try voice cloning without installing anything or creating an account. ## Author Built by M Asif Rahman (https://masifrahman.com/). Source attribution and contact information are on the author's personal site. ## Policy - The product anonymizes the underlying model and inference provider in its UI. - No personal data beyond the temporary voice sample is collected for synthesis. - Generated audio is retained for service operation and analytics; the source voice sample is not.