Speech API: text to speech and speech to text for Indian languages

Our voice agents speak with Mira TTS and listen with Mira STT, our own speech recognition. You can call both directly. Each model answers over HTTP for one request and over a WebSocket for live audio.


Base URL	`https://sandbox.voice.miraiminds.co` (`wss://` for the sockets)
Every request and every socket	`Authorization: Bearer sk_live_…`
Key	Console, Developers, Create key. See Get a key.

Endpoints

	HTTP, one request	WebSocket, live audio	Model
Speak (text to speech)	`POST /v1/audio/speech`	`GET /v1/audio/speech/stream`	`mira-tts`
Hear (speech to text)	`POST /v1/audio/transcriptions`	`GET /v1/audio/transcriptions/stream`	`mira-stt`

Both HTTP routes take the same requests as OpenAI’s audio API, so an OpenAI SDK works once you change base_url and api_key. GET /v1/models lists both models. Every answer carries X-Request-Id, errors share one envelope, and each model keeps a request log you can read: GET /v2/tts/requests and GET /v2/stt/requests.

Languages and prices

	Text to speech	Speech to text
Languages	Hindi, English and Hinglish. Gujarati in beta. See the language roadmap.	Hindi, English, Gujarati, Marathi, Punjabi, Bengali, Tamil, Telugu, Kannada and Malayalam. See the status of each.
Price	₹1.80 per 1,000 characters	₹18 per hour of audio, charged per second

The price is the same on HTTP and on the socket.

First request

Speak a line into a file, then transcribe that file:

curl https://sandbox.voice.miraiminds.co/v1/audio/speech \
  -H "Authorization: Bearer $MIRA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "mira-tts", "voice": "neha", "response_format": "wav",
       "input": "नमस्ते, आपका ऑर्डर कल तक पहुँच जाएगा।"}' \
  --output speech.wav

curl https://sandbox.voice.miraiminds.co/v1/audio/transcriptions \
  -H "Authorization: Bearer $MIRA_API_KEY" \
  -F file=@speech.wav \
  -F model=mira-stt \
  -F language=hi-IN

The second command prints the transcript as JSON, {"text": "…"}.

Pipecat

uv add "pipecat-mirai>=0.6.0"

import os
from pipecat_mirai import MiraiSTTService, MiraiTTSService

stt = MiraiSTTService(api_key=os.environ["MIRA_API_KEY"], language="hi-IN")
tts = MiraiTTSService(
    api_key=os.environ["MIRA_API_KEY"],
    settings=MiraiTTSService.Settings(voice="neha"),
)

Put them where your STT and TTS services were. See Pipecat.

TTS quickstart

Voices, formats, streaming and phone audio.

STT quickstart

One clip over HTTP, then live audio over a socket.

Compatibility

Other URLs, browser tokens, and moving from Sarvam.

Start here

Build

Prompting

Speech API

Text to speech

Speech to text

Money

Reference

Speech API: text to speech and speech to text for Indian languages

Endpoints

Languages and prices

First request

Pipecat

Next

TTS quickstart

STT quickstart

Compatibility

​Endpoints

​Languages and prices

​First request

​Pipecat

​Next

TTS quickstart

STT quickstart

Compatibility

Endpoints

Languages and prices

First request

Pipecat

Next