Voice Studio is a voice-first creative app. Speak to control brushes, paper, colors, and
physics; generate faint pencil sketch guides; and get quick AI replies that mirror your
language style (including Roman Urdu / Hinglish). The goal: zero friction from idea to
stroke.

What you can do

- Voice control: “Brush size 8,” “more water,” “lower ink to 30,” “switch to watercolor
paper,” “undo,” “redo,” “clear,” “toggle view.”
- AI sketch guides: Say “faint pencil outline of a skyline / portrait / product” and paint
over the generated guide.
- Real-time physics: Brush size, water, ink, drying speed, viscosity, paper resist,
pigment weight, brush types (round/flat/sumi/spray/water).
- Paper + color: Choose paper type (rice, canvas, watercolor, smooth, parchment, textured
matte), roughness, contrast, resolution; set hue/saturation/brightness by voice.
- Transcript history: Recent voice exchanges saved locally and viewable in the History
tab.
- Language match: Replies mirror your latest style; switch to English on request.
- Barge-in: If you start speaking, AI speech stops immediately and listens.
- Performance: Low thinking budget, continuous audio streaming for stable turn detection,
echo cancellation, noise suppression, auto gain, single-channel capture, remembered mic
device ID.
- Reliability: Sketch generation uses fallback models and timeouts to avoid hangs; clear
errors if a model is unavailable.
- Context recall: Recent conversation + last sketch prompt injected into AI context each
session to reduce “forgetfulness.”

Use cases

- Rapid ideation by voice while keeping hands on stylus
- Guided practice (faces, anatomy, grids, lettering) with faint outlines
- Live teaching/demos—speak adjustments while students watch changes
- Palette/material shifts in seconds; experiment with watercolor-like flows

Tips for best results

- Stable internet for fastest responses
- Short, specific prompts for sketch guides
- Speak clearly in a quiet space or use a headset mic
- Interrupt anytime by speaking; AI stops and listens

Data, privacy, permissions

- Microphone: required for voice features, used while you’re interacting
- Transcripts: recent history stored locally on your device for convenience; not sold or
shared
- Network: needed for AI processing; no offline mode yet
- Safety: If speech is unclear, the app asks you to repeat; timeouts prevent hangs