TTS Studio
ⓘHardware Requirements
Best on Apple M-series (M1/M2/M3/M4) using Metal GPU acceleration.
Intel Macs are supported but run on CPU only — expect
5–10× slower generation.
Windows/Linux with an NVIDIA GPU also works great via CUDA.
Text-to-speech studio for long-form audio.
Model Status: Idle
A new version () is available!
Saved
Voice
Tell the model how to deliver the text — pace, tone, emotion, energy. Leave blank for the speaker's default style.
Low (0.1–0.3): consistent, controlled delivery. High (0.7–1.0): more expressive and varied, but less predictable between runs.
Input Text
Paste Markdown (not plain text) so headings come through. In Google Docs: Tools → Preferences → Automatically detect Markdown, then copy. Text is split into paragraphs automatically; each ## heading becomes a chapter start.
Paragraphs (0)
🔑 Enter Access Code
This TTS Studio server is private. Paste the access code you were given — you'll only need to do this once on this device.
That code wasn't accepted — check it and try again.
📚 How to Use TTS Studio
1
Pick a voice
In the Voice panel, choose a Mode: Voice Cloning (clone any
voice from a short audio sample), Preprogrammed Voice (built-in
speakers), or Voice Design (describe the voice you want in words).
The 1.7B model sounds best; 0.6B is faster.
2
Paste your text
Paste Markdown into the Input Text box and click
Parse Paragraphs. Your text is split into paragraph cards, and each
## heading becomes a chapter start. The first line becomes the project
name.
3
Generate audio
Click Generate All (or ⌘+Enter) to synthesize every
paragraph, or use each card's Generate button one at a time. The
first run downloads the model, which can take a few minutes.
4
Review and fix
Play each card's audio. Don't like a paragraph? Edit its text right in the card or hit
Regenerate — each attempt is kept as a take so you can
A/B them and keep the best. You can also reorder, insert, or delete paragraphs.
5
Export
When every paragraph shows Ready, choose WAV or
M4A and click Download Audio to get one merged file.
Your work auto-saves as a project — reopen it anytime via
Open in the project bar. Press Esc to close any window like this
one.
Projects
No saved projects yet.
Save New Voice Profile
Activity Log
⚙ Settings & Diagnostics
Generate audio on a shared server instead of this
machine — useful if your Mac is slow. Only synthesis runs remotely; your projects,
voice profiles, and exported audio all stay local. Leave the URL empty to generate
locally.