Introducing VocaSync: A Voice Platform for Creators and Developers
I've been busy for a while now, working on a product I've wanted to ship since the last time I thought Valeon had reached it's feature-complete state. If I were completely honest with myself, it's been a bit of an arduous journey, I've built, deleted and rebuilt this project at least 3 different times, designing, and redesigning not just the interface, but more importantly the architecture underneath. It's not that my propensity for perfectionism took over me, but the vision I had for a product that could scale had to be done in a certain way, so this is what I have.
Today we're introducing VocaSync by Valeon: a focused voice platform for creators and developers who want synthetic speech and precise subtitles to feel like first-class output, not an afterthought. VocaSync provides a simple dashboard when you want to move fast, and a clean REST API when you want it embedded into a workflow.
It's built around one idea: voice should be reliable, repeatable, and easy to orchestrate, whether you're publishing narrated essays, shipping an accessibility feature, or generating synchronised media at scale. It's also designed from the ground up to be scalable—powered by modular workers—so it can grow from a single job to a high-throughput pipeline without changing the way you integrate it.
At launch, VocaSync brings two capabilities under one roof: speech synthesis and forced alignment. On the synthesis side, you can generate natural speech with nine distinct voices—alloy, ash, coral, echo, fable, onyx, nova, sage, shimmer—and export in MP3, AAC, OPUS, FLAC, or WAV, with options geared for quality and low-latency delivery.
On the alignment side, you can synchronise existing audio with its transcript to produce word-level timestamps and export subtitle formats like SRT and VTT, with support for English (US), English (UK), French, German, Spanish, Russian, and Ukrainian—each backed by language-specific modelling to keep timing crisp even as content gets long and messy. This is not "audio generation" as a novelty, but voice as infrastructure: the kind you can depend on when your output needs to ship.
VocaSync is built to plug into automation just as easily as it serves a human workflow: programmatic job creation, predictable artefacts, and an architecture that treats voice processing as a repeatable service you can compose into your stack. VocaSync is also designed to be straightforward to adopt: pay-as-you-go, no subscriptions, with a free monthly allowance so you can test real workloads before you commit.
Synthesis starts at 3p per 1,000 characters, alignment at 2p per minute of audio, and every account includes 10 minutes of alignment plus 1,000 characters of synthesis per month. If you've ever wanted to turn writing into listenable media, make your product more accessible, or ship multilingual voice features without building an entire audio stack from scratch, VocaSync is the on-ramp.