Kokoro Voice Studio gives you 50+ natural voices, a real-time voice mixer, one-click emotion presets and pro audio post-processing — with zero servers and zero uploads. Your text and audio never leave your device.
From quick voiceovers to long-form audiobooks — Kokoro Voice Studio packs a full production chain into one focused interface.
Curated, human-sounding voices spanning male, female and non-binary registers. Each voice is sample-accurate at 24 kHz.
Blend any two voices with a single slider — get a brand-new voice in under a second. Save your blends to your private library.
Cartoon, Hero, Whisper, Happy, Sad, Polish — every preset auto-renders the moment you tap it. Fine-tune intensity with a built-in slider.
Warmth, de-essing, voice doubling, room tone, multiband compression and LUFS normalisation — broadcast-quality finish in a click.
Powered by ONNX Runtime Web with WebGPU acceleration. No upload, no cloud, no telemetry — your scripts and audio never leave your machine.
Install once, generate forever — even on a flight. The app caches its model and voices locally and works fully offline after the first visit.
Paste any text — a sentence, a chapter, a podcast intro. Inline markup like [pause] and {emphasis} is supported.
Choose from 50+ voices or blend two together with the mixer. Save your favourite blends to your private voice library.
Select an emotion preset, tweak the intensity slider and toggle Pro DSP — every change auto-renders the audio in real time.
Download as 24-bit WAV, 320 kbps MP3 or lossless FLAC. Use it in your video, podcast, game or app — royalty-free.
Start free in your browser. Upgrade only when you need cloud sync, longer documents or a self-hosted edition.
Everything an indie creator needs to ship beautiful voiceover from a laptop.
For podcasters, indie game devs and content teams who want sync, scale and the full clone toolkit.
For studios, agencies and enterprises with on-prem or sovereignty requirements.
All plans include royalty-free commercial use of generated audio. GST extra where applicable.
CodeKins Pvt. Ltd. is an independent software studio crafting on-device AI tools for creators. We believe powerful media tools shouldn't require accounts, subscriptions or cloud uploads to be useful.
Kokoro Voice Studio is our flagship product — a love-letter to the open-source Kokoro 82M speech model, wrapped in a producer-grade interface that runs entirely inside your browser.
No. Kokoro Voice Studio runs entirely in your browser using ONNX Runtime Web. The model and voice files are downloaded once from a public CDN and cached locally. After that, every generation, mix and effect happens on your device — no requests are made to our servers.
Yes. Audio you generate with Kokoro Voice Studio is royalty-free for both personal and commercial use, including ads, podcasts, games, films and audiobooks. You don't need to credit us, but a mention is always appreciated.
Any modern Chromium-based browser (Chrome, Edge, Brave, Arc) gives you the fastest experience via WebGPU. Safari and Firefox work too with the WebAssembly fallback — slightly slower on first generation but indistinguishable in audio quality.
About 83 MB for the default Balanced quality tier — downloaded once and cached forever. Lite mode (~42 MB) is available for slow connections; Studio FP16 (~155 MB) is offered for users who want zero compromise on fidelity.
No accounts, no API keys, no per-character charges, no rate limits, no waiting on a network round-trip and — most importantly — your scripts never touch our servers. The trade-off is the one-time model download; after that you have an unlimited TTS engine in your pocket.
Yes — Kokoro Voice Studio is a Progressive Web App. In Chrome, click the install icon in the address bar; in Safari on iOS/iPadOS, use Share → Add to Home Screen. It then runs in its own window, fully offline.
The Studio loads in under three seconds — type a sentence, pick a voice, hit generate. No sign-up.
Launch Kokoro Voice StudioSales enquiries, partnership ideas, support questions — pick whichever channel feels easiest.