DocumentationQuickstart
Models
Oogam exposes DWET's voice models behind stable model IDs. Pass the ID in the model field of your request. Speech-to-text requires a languagecode (our models don't auto-detect).
| Model ID | Capability | Pricing | Description |
|---|---|---|---|
naad-tts-v1 | Text to Speech | ₹0.30 / 1k characters | Expressive, low-latency text-to-speech across 22 Indian languages. |
naad-tts-v1-turbo | Text to Speech | ₹0.18 / 1k characters | Fastest TTS for real-time use; slightly lighter prosody. |
naad-stt-v1 | Speech to Text | ₹0.60 / audio-minute | Speech recognition + translation for 22 Indian languages, code-mixed and telephony audio. |
naad-s2s-v1 | Voice Changer | ₹0.90 / audio-minute | Re-voice any audio into a target voice while preserving content and timing. |
naad-sts-v1 | Speech to Speech | ₹0.75 / audio-minute | Speech-to-speech — speak or upload audio and hear it in a target voice. Live talk uses the NAAD realtime API. |
naad-isolate-v1 | Voice Isolator | ₹0.45 / audio-minute | Remove background noise and isolate clean speech from any recording. |
More products (chat, code, vision and agents) are on the roadmap and will be listed here once they ship.
Try the voice models live in the Voice Studio.