API reference index

Send HTTP TTS requests to the model’s host, such as https://coda.api.rime.ai/. For Coda WebSocket v1, connect to wss://coda.api.rime.ai/ws. Authenticate with a bearer token in the Authorization header. See API authentication. Most clients should start with Coda, Rime’s flagship model:

• Coda: Streaming HTTP
• Coda: WebSocket v1 reference
• Coda: WebSocket v1 JSON quickstart
• Coda: WebSocket v1 Protobuf quickstart
• Coda: WebSocket v1 with LiveKit

To compare model and protocol support, see the WebSocket API overview.

TTS endpoints by model

Mist v1 and Mist v2 retire on December 3, 2026. To move from Mist v1 or Mist v2, see Move from Mist v1 or Mist v2 to Mist v3.

Speech-to-text endpoints

Transcribe English or Spanish with your Rime API key. Both endpoints accept streaming 16 kHz mono PCM16 audio. gRPC also accepts a complete WAV file.

• Speech-to-text: Streaming WebSocket: GET wss://stt.api.rime.ai/ws
• Speech-to-text: gRPC: stt.api.rime.ai:443, service rime.SpeechToText

To transcribe from Python, Node.js, Go, or Rust, start with the speech-to-text quickstart, which uses the Rime SDK. To call the API without the SDK, Call the API directly helps you choose a transport and an output contract.

Prism endpoint

Hold a spoken conversation with Prism, Rime’s speech-to-speech API, over one WebSocket.

• Prism: WebSocket: GET wss://prism.api.rime.ai/v1/realtime

To build with the Rime SDK, start with the Prism SDK quickstart. To call the API without the SDK, follow the WebSocket quickstart.

Metadata endpoints

• List All Voices: GET /data/voices/all-v2.json
• List Voice Details: GET /data/voices/voice_details.json

Utility endpoints

• Coverage (out-of-vocabulary check): POST /oov
• Text Normalization: POST /textnorm

Related

• Quickstart: TTS in five minutes
• Models: comparing Coda and Mist
• Voices: voice lineup and language support
• Latency tuning
• Changelog