What Tablif read on Hlas-1's site
Hlas-1 describes itself as: “Hlas-1 Expressive, controllable speech in a 3.5B open-weight model”
- Price model: free tier
- Stage: Live, no waitlist
What it is
Calls itself “text-to-speech model”.
- Calls itself: text-to-speech model
“…text-to-speech model”
What it costs
Has a free plan.
- Free plan: yes
“Free for research and non-commercial use.”
Who it sells to
In multilingual.
- Languages: multilingual
“…generate multilingual speech.”
How it runs
A web app. Runs locally and works offline. Open source.
- Kind: web
- Runs where: on the device, offline
- Open source: yes
“The voice stays in this browser”
“The voice stays in this browser and is sent only when you generate with it.”
What it works with
Names Breeze TTS 2 and Hugging Face on its pages.
- Names: Breeze TTS 2 · Hugging Face
“Built as a spare-time project on Breeze TTS 2”
“Hlas-1 is a 3.5B open-weight text-to-speech model”
What it offers as proof
Live, no waitlist. Try it without signing up.
- Trial asks for: nothing
“Model weights (opens Hugging Face in a new tab) Home”
“The voice stays in this browser and is sent only when you generate with it.”
Sources
Tablif read these pages, last on 22 Nov 2026. Everything above is what the product says about itself; Tablif does not verify every claim.
Similar products in Tablif Town
- SpeechWave: Expressive AI voiceovers that direct themselves, from $3/mo
- Voxelle: Free AI voice generator: turn text into natural speech and d
- Murmur: Free, native voice-to-text for macOS
- DessaVoice: Zero-text speech-to-speech, explained
- Dictaro: Speech to text in your browser, with AI rewrite
- Earlybird: 29M-param speech-to-text that outputs words before they end
- Lowspeak: AI speech-to-text in every Mac app, online or offline
- VoiceHue: Free multilingual text-to-speech with 138 natural voices