Processing, please wait...
Do not refresh the page
Voice Management
Manage voice samples, global NPC fallback voices, and pronunciations across all TTS providers.
Note: XTTS, Chatterbox, and PocketTTS share a simple voice sample flow. OmniVoice imports voices into the selected language library.
Fallback Voices
Choose the voice used when an NPC has no explicit voice assigned. These settings apply to every TTS connector.
Resolution order: explicit NPC voice, matching race and gender voice, then the connector's male or female fallback. Leave a field blank to skip the race fallback for that combination.
Pronunciations
Rewrite how the TTS engine says a word without changing anything the player reads. These entries apply to every TTS connector.
- Audio only: subtitles and saved dialogue keep the original spelling — only the text sent to the voice engine is rewritten.
- Blank field: that filter is not applied. With NPC names, races, and Oghma tags all blank the entry is global and every NPC uses it.
- Commas inside one field are alternatives — Nord, Dunmer matches either race.
- Two or more fields filled: the speaking NPC must match all of them, so Nord plus companions only fires for a Nord carrying that Oghma tag.
- Built-in entries keep their original term, but their spoken version can be edited and they can be disabled or deleted.
Preview voice
Pick a connector and voice, then use a play button beside any entry.
Add Custom Pronunciation
Original text is matched as a whole term. Leave every access field blank to apply the entry to all NPCs.
Custom Pronunciations
0 custom entries.
No custom pronunciations yet. Add one above to override how a word is spoken.
Built-in Pronunciations
Voice Sample Upload
Upload voice samples to data/voices. Files will be cached and uploaded to the XTTS server.
XTTS Info: XTTS is a TTS engine that generates cloned voices from samples. Uses roughly 4GB of VRAM. Best for NVIDIA GPUs.
📋 File Requirements:
- Format: WAV (PCM), 16-bit, Mono, 20500Hz
- Size: 5MB or less
- Filename: lowercase with underscores (e.g., "mjoll_the_lioness.wav")
- XTTS: Audio clips must be at least 5 seconds long for voice generation
Note: If replacing an existing voice, restart the XTTS server after upload.
XTTS Voice Cache
Manage voice samples for XTTS. Voices are uploaded from local .wav files in data/voices to the XTTS server.
Batch Process Missing Voices
Upload missing local voices to the XTTS server.
Found 1 voice(s) not yet uploaded to XTTS server.
Cloud XTTS Sync
Advanced XTTS configuration: http://127.0.0.1:8020/docs
Voice Sample Upload
Upload voice samples to data/voices. Files will be cached and uploaded to the Chatterbox server.
Chatterbox Info: Chatterbox is an optimized fork of XTTS with faster inference. Uses roughly 4GB of VRAM. Shares the same API as XTTS.
📋 File Requirements:
- Format: WAV (PCM), 16-bit, Mono, 20500Hz
- Size: 5MB or less
- Filename: lowercase with underscores (e.g., "mjoll_the_lioness.wav")
- Chatterbox: Audio clips must be at least 5 seconds long for voice generation
Note: If replacing an existing voice, restart the Chatterbox server after upload.
Chatterbox Voice Cache
Manage voice samples for Chatterbox. This view now shows all voices currently reported by the live Chatterbox endpoint, plus any local cached samples in data/voices.
Batch Process Missing Voices
Upload missing local voices to the Chatterbox server.
Found 1 voice(s) not yet uploaded to Chatterbox server.
Cloud Chatterbox Sync
Advanced Chatterbox configuration: http://127.0.0.1:8023/docs
Voice Sample Upload
Upload voice samples to data/voices. TTS Studio automatically uses local files for audio.cpp or server synchronization for the standard API.
Detected PocketTTS Mode: audio.cpp at http://127.0.0.1:8086.
📋 File Requirements:
- Format: WAV (PCM), 16-bit, Mono, 20500Hz
- Size: 5MB or less
- Filename: lowercase with underscores (e.g., "mjoll_the_lioness.wav")
- PocketTTS: Audio clips must be at least 5 seconds long for voice generation
Note: If replacing an existing voice, restart the PocketTTS server after upload.
PocketTTS Voice Cache
Manage voice samples for PocketTTS. Current mode: audio.cpp.
With audio.cpp PocketTTS on another host, uploads stay on this mod server and are sent with each request, like Higgs. For audio.cpp on a custom port, set the connector endpoint to the full /v1/audio/speech URL.
Batch Process Missing Voices
Make missing local voices available to PocketTTS.
✓ All local voices are available to PocketTTS.
Cloud PocketTTS Sync
Advanced PocketTTS configuration: http://127.0.0.1:8086/docs
Higgs TTS 3 Voices
Upload a WAV sample, then preview its voice. Previewing does not change your active TTS connector.
Uploads stay on this mod server and are sent with each request to a remote Higgs service. Named voices on that host also appear here after refresh.
No Higgs voice samples found. Upload a sample or check your remote voice library.
OmniVoice Language Library
Choose the language library that uploads, sync, and tests should use.
Voice Sample Upload
Upload WAV samples into the selected OmniVoice language library. Reference text is generated automatically by local STT.
File Requirements:
- Format: WAV voice reference sample
- Filename becomes the VoiceID, such as
femalenord.wav - No transcript file is required; local STT creates the reference text
OmniVoice Voice Library
Manage voices for the selected language library: en.
Batch Import Missing Voices
Import missing local voice samples into the selected OmniVoice language library.
Found 1 voice(s) not yet ready in OmniVoice for en.
OmniVoice Service
Advanced OmniVoice configuration: http://127.0.0.1:8021/docs
Voice Sample Upload
Upload voice samples to data/voices. Files will be available for generating voices in Cartesia.
📋 File Requirements:
- Format: WAV (PCM), 16-bit, Mono, 20500Hz
- Size: 5MB or less
- Filename: lowercase with underscores (e.g., "mjoll_the_lioness.wav")
Note: Files will be saved to data/voices and automatically generated for Cartesia.
Cartesia Voice Cacje
Manage voice generation for Cartesia TTS. Voices are generated from local .wav files in data/voices.
For detailed information, see our Cartesia TTS Guide.
⚠️ Cartesia API not configured
Please configure your Cartesia API key in the API Badge page before syncing voices.
ℹ️ Automatic Voice Generation
Voices are automatically generated when you speak to an NPC using that voice in-game, if Cartesia TTS is selected as your TTS provider. You don't need to manually sync all voices upfront.
Batch Generate Missing Voices
Generate voice clones for Cartesia from uploaded files. Rate limited to avoid API errors (2 second delay between requests).
Found 1 voice(s) not yet generated for Cartesia.
Estimated time: ~1 minute(s)
Voice Sample Upload
Upload voice samples to data/voices. Files will be available for generating voices in Inworld.
📋 File Requirements:
- Format: WAV (PCM), 16-bit, Mono, 20500Hz
- Size: 5MB or less
- Filename: lowercase with underscores (e.g., "mjoll_the_lioness.wav")
Note: Files will be saved to data/voices and automatically generated for Inworld.
Inworld Voice Cache
Manage voice generation for Inworld TTS. Voices are generated from local .wav files in data/voices.
For detailed information, see our Inworld TTS Guide.
⚠️ Inworld is not fully configured
Please configure an active Inworld connector and API credential and workspace before syncing voices. Set the API credential in API Badge and the workspace on your active Inworld TTS connector.
ℹ️ Automatic Voice Generation
Voices are automatically generated when you speak to an NPC using that voice in-game, if Inworld TTS is selected as your TTS provider. You don't need to manually sync all voices upfront.
Batch Generate Missing Voices
Generate voice clones for Inworld from uploaded files. Rate limited to avoid API errors (3 second delay between requests).
Found 1 voice(s) not yet generated for Inworld.
Estimated time: ~1 minute(s)