SpeakEasy

Video Demo

About this project

SpeakEasy is a browser-based, real-time speech coach for rehearsing any talk—pitches, presentations, interviews, class projects—so you walk in confident. Most people learn how they really sound only after it's too late. SpeakEasy gives you that feedback while you can still use it. Before you start, paste your material and describe your audience and setting. As you speak, SpeakEasy listens live and tracks what makes or breaks delivery: pace, volume, pitch variation, pauses and dead air, and filler words like "um" and "uh." A calm on-screen nudge surfaces one fix at a time (slow down, project, add energy), so you're guided, never overwhelmed. Want your eyes off the screen? Turn on optional spoken cues that coach you out loud, hands-free. When you finish, SpeakEasy turns the session into a detailed report: real delivery metrics, a pace timeline across your whole talk, and a full transcript with your naturally emphasized words highlighted. Google's Gemini reads your actual material alongside how you delivered it to produce advice tuned to your specific audience, plus tone-content mismatch detection that flags moments where how you sounded didn't match what you said. Every session saves to your private history, so you can track progress over time. It runs entirely in the browser with nothing to install, built on Google Cloud end to end: Gemini on Vertex AI for the report and tone analysis, Cloud Speech-to-Text for transcription, Firebase Authentication for Google sign-in, Cloud Firestore for per-user history, and Cloud Run behind Firebase Hosting for the API. AI keys stay server-side and never reach your browser. Whether you're prepping for a big moment or sharpening everyday communication, SpeakEasy gives you a private room to practice, honest real-time feedback, and an AI coach that understands not just how you speak, but what you're trying to say.

Gallery