Introducing voice finder — a new tool to quickly find the right voice for your app from over 600+ voices
Together AI
Together AI shipped Voice Finder, a search tool for picking AI voices out of a giant pile. Describe what you want or upload a sample, and it matches you against 600+ TTS voices instantly.
Based on reporting by Together AI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Picking a voice for a text-to-speech app has always been a weirdly tedious job. You end up scrolling through dozens, sometimes hundreds, of samples, clicking play over and over, trying to remember which one sounded warm enough or fast enough or not too robotic. Together AI apparently got tired of watching developers do this by hand, so they built Voice Finder, a search layer sitting on top of their text-to-speech models.
The pitch is simple: instead of browsing a static list, you type what you're after in plain English — something like 'calm female narrator, mid-30s, slight British accent' — and the tool surfaces matches from a library of more than 600 voices spread across Together's TTS lineup. You can also upload an audio clip and ask it to find voices that sound similar, which is handy if you're trying to match an existing brand voice or replace a narrator without starting from zero.
Beyond search, Voice Finder lets you filter results and audition candidates directly, so the whole discovery-to-decision loop happens in one place rather than bouncing between docs, sample pages, and API calls. That sounds minor until you've actually done this work manually and realized how much time gets burned just narrowing down options before you even touch code.
It's a small feature in the grand scheme of Together's product lineup, but it points at something bigger: as TTS catalogs balloon into the hundreds of voices, the bottleneck stops being model quality and starts being findability. Together is betting that better search, not more voices, is what actually unlocks adoption for teams building voice agents, audiobooks, or customer-support bots.
My take — AI-written commentary, not fact-checked reporting
Honestly, this is the kind of unglamorous tooling that matters more than another benchmark chart. Voice selection has been an underrated pain point for years, and search-by-description plus audio matching is the obvious fix nobody bothered shipping. My only gripe is it ties you deeper into Together's own model garden rather than working across providers — convenient today, a lock-in headache the moment you want to switch.
Read more about this at: Together AI