Text to speech
Piper
CPU-friendly English TTS with a pinned, predictable voice.
- Status
- Live
- Languages
- English on the current alias
- License
- GPL-3.0 runtime; voice licenses vary
Model ID
| Model ID | Designed for |
|---|---|
piper | Efficient English speech with a fixed voice |
Confirm access with GET /v1/models, then follow the utility-model docs for the canonical v1 REST or native gRPC request contract.
Pricing
piper: Free for a limited time. $0.00005 $0.00 per audio second
Overview
Piper is a compact, CPU-friendly speech engine. Tarka currently exposes one
pinned English voice through the piper alias for efficient prompts,
notifications, and test audio.
When to pick it
- Volume is high enough that CPU efficiency dominates
- The output is going to be heard once and discarded (notifications, one-off briefings)
When to pick something else
- Voice cloning → Qwen3-TTS
- Named English voices → Kokoro
- Nepali speech → Nepali Parler TTS
Limits
Send voice: “default” to POST /v1/audio/speech or
the native gRPC speech method. Use model discovery and your organization limits
as the source of truth for access and request limits.