A vibe-designed cute voice assistant

🚀 Vibe-designed and coded a cute voice assistant using Replit, Midjourney, ElevenLabs & OpenAI Whisper!

𝗪𝗼𝗿𝗸𝗳𝗹𝗼𝘄 𝗯𝗿𝗲𝗮𝗸𝗱𝗼𝘄𝗻:

 1. Built the design strategy with deep ChatGPT research

 2. Designed the character in Figma + GPT Sora

 3. Animated states with Midjourney

 4. Crafted unique voice styling via ElevenLabs

 5. Used OpenAI Whisper for seamless speech-to-text integration

𝗞𝗲𝘆 𝗶𝗻𝘁𝗲𝗿𝗮𝗰𝘁𝗶𝗼𝗻 𝗱𝗲𝘀𝗶𝗴𝗻 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀

 • Four dynamic voice states—idle, listening, thinking, and speaking—each animated in real time by user input

 • Synchronized voice playback with word-by-word text highlighting

 • Seamless transitions for a lively, responsive character

 • Unified prompt and animation language for consistent personality

Back when building Miko robot, building a similar voice experience would easily took 2–3 weeks. This time: just 3 days for planning, building, and debugging.

AI workflows are a game changer