Built in Athens.
For every language.
Orpheus is a studio-grade text-to-speech service designed around two ideas: Greek should sound like Greek, and 600+ other languages shouldn't feel like afterthoughts. With Pro, you can also clone your own voice and narrate in any of them.
Who makes it
Orpheus is a product of Multimodal Solutions, a small Athens-based studio that ships custom software, macOS apps, and AI products. Our shipped work includes MacPulse, Captain's Deck, NostalgiApp, and others — all independently funded, all running in production.
Why Orpheus
Most open-source TTS pipelines are fantastic at English and Spanish and progressively weirder the further you get from their training data. Greek in particular — the language we speak every day — is usually an afterthought in these catalogs: single voice, occasional pronunciation gaps, flat prosody.
We went the other way. Every language gets matched to the engine that sounds most natural for it: our own GPU stack for English and Spanish, and commercial studio-grade voices from Google Cloud for the rest, including Greek. We insert the punctuation pauses a native speaker would take. We hide voice slots that don't actually exist for a given language instead of pretending they do.
The goal is simple: a paragraph of Greek, Hungarian, or Hindi should land in your podcast, audiobook, or client deck feeling like a real person read it — not a robot pretending to.
What we promise
- 600+ languages, with the 18 most-asked for in the UI today.
- Commercial use included. No watermarks, no forced attribution.
- Pay only for the days you use it — €4.99 Day Pass, €9.99 Pro with voice cloning. No auto-renewal.
- Your text and your voice clones never feed model training. We don't sell data.
- Clear pricing, clear terms, clear consent rules for voice cloning.
Say hello
Feature request? Weirdness in a language you speak? Business question? Email us at [email protected]. We read everything.