Text-to-Speech Providers
This is the reference for configuring Read Aloud voices. For playback in the viewer, see Read Aloud.
"Read Aloud (TTS)" is its own entry in Settings (a full screen on mobile, a dialog on desktop) and is also reachable from the Read Aloud sheet. Pick a provider from the row of cards; the panel below swaps to match. Audio plays through the shared platform audio player.
Desktop
MobileProviders
| Provider | Tier | Notes |
|---|---|---|
| System / Device | Free | The device's built-in voice |
| ElevenLabs | Pro | Premium AI voice (cloud) |
| OpenAI / Gateway | Pro | OpenAI / OpenRouter / custom |
| Local AI (Kokoro) | Pro | On-device neural (desktop) |
System / Device
Pick the device voice; Test plays a preview.
ElevenLabs Pro
Enter your API key (with show/hide and Validate), then choose a voice and Test.
OpenAI / Gateway Pro
A preset dropdown (OpenAI / OpenRouter / custom) pre-fills the endpoint. Set your API key (with Validate), model, voice, and output format.
Local AI (Kokoro) Pro
A self-hosted, OpenAI-compatible speech server (desktop):
- Set the local inference endpoint (default
http://localhost:8880/v1/audio/speech) and use Check Daemon to verify it. - Registered voice models are listed with Set Active / Test / delete.
- Add Local Model / Voice opens a form (identifier + locale + gender). Voice
ids are server-specific (for example
af_bella,am_adam).
Playback features
- Background playback Premium, speed 0.5×–2.0×, auto-scroll, provider fallback, and per-page or continuous reading. Details in Read Aloud.