I started by messing around with text-to-speech - swapping voices, testing models, trying to make synthetic speech sound less stiff and more useful. One tool for a script. Another for a PDF. Another if you wanted your own voice. It was fun to tinker with, and annoying to use. So AnyToSpeech became the workshop I wanted: convert text, documents, and images to speech, clone a voice, transcribe audio, and generate podcasts without hopping between apps just to try the next model. The same curiosity spilled into the more entertaining side of voice AI - singing ratings, karaoke-style practice, accent checks. Just proof that voice models can be useful and entertaining in the same product. If you buy the deal and need help, email me at [email protected]. I read every message and reply myself as soon as I can.