Design goal
Automatic running focuses attention on jump height and landing time. Voice is not a decorative switch but continuous input; players form a rhythm between sound and silence that feels distinct from an immediate key press.
Space and touch remain available so voice never becomes an entry barrier. Every input uses the same physics, terrain, shields, enemies, and scoring rules.
Technology and art
React manages page state, microphone access, pausing, leaderboard, and local progress. Phaser Arcade Physics handles movement, collision, enemy patrols, and streamed terrain. Cloudflare Pages Functions and D1 provide run tickets, score validation, and the anonymous leaderboard.
The moonlit paper-cut forest, runner, enemies, mist, obstacles, and particles are built from procedural graphics. This keeps loading light and gives each act a distinct tone within one visual language.
AI-assisted development
ChatGPT, OpenAI Codex, and GPT assisted with prototyping, code collaboration, and documentation. Product goals, level tradeoffs, final review, deployment, and responsibility for published content remain with the project maintainer.
Disclosing AI assistance does not make generated material the product value. SoundLeap's value comes from an original playable mechanic, tuned level data, collision rules, score validation, and ongoing maintenance.
Feedback is open
The source code is publicly hosted on GitHub. Device compatibility, accessibility, level defects, and privacy concerns can be reported through repository issues with browser, operating system, and reproduction steps.
Turn your voice into wind beneath your feet
Play now →