hjLabs AI Playground
Starting the on-device runtime…
Kokoro-82M is a remarkably small neural TTS model that sounds far better than its 82 million parameters suggest, and it runs comfortably in a browser tab. Pick a voice, paste text, and get a WAV you can download — no character quota, no per-thousand-character billing, and no upload of the script. SpeechT5 and MMS-TTS are available as alternatives, MMS in particular for languages Kokoro does not cover.
free text to speechkokoro tts onlineoffline tts browserai voice generator freetts no character limitdownload tts wavspeecht5 demolocal voice synthesis
Check the model licence on Hugging Face — Kokoro-82M is Apache-2.0, which permits commercial use. We add no additional restriction of our own; the audio is generated on your machine and is yours.
That is the one-time model load and GPU warm-up. Subsequent generations are much faster because the weights are already resident.
Yes, once a model is cached. Weights are stored in the Cache Storage API under the key "transformers-cache", so a reload with no network still runs the model. Call "Make storage persistent" first — browsers may evict non-persistent caches under disk pressure.
No. The model weights are downloaded from Hugging Face to your browser once, and every inference after that runs on your own GPU through WebGPU. Your text, images and audio are never sent anywhere — hjLabs.in has no inference server and no way to see your input.
Small models do. Recent mobile Chrome and Safari expose WebGPU, but memory is the binding limit — stay under roughly 200 MB of weights on Android, and expect iOS to reclaim the tab sooner. The catalog shows every download size up front so you can pick something that fits.
That depends on the individual model, not on this site. Each model page links to its Hugging Face repository, where the licence is stated — most of the catalog is Apache-2.0 or MIT, a few are research-only. We add nothing on top: hjLabs.in claims no rights over anything you produce here.