Integration · Voice infrastructure

Oogam + LiveKit

LiveKit handles real-time audio, rooms, telephony and SIP. Add Oogam's speech models to build voice agents that listen and speak in 27 Indian languages and English — speech to text, text to speech and Sutra as LiveKit Agents plugins.

What's included

Three plugins, one package.

Speech to Text

oogam.STT

Streaming transcription in the caller's language. Finalizes each turn the moment LiveKit hears the caller stop, and can follow a caller who switches language mid-call — most reliably from Hindi to English today.

wss://platform.oogam.ai/v1/realtime?intent=transcriptionReference

Text to Speech

oogam.TTS

Speech streamed as it is generated, sentence by sentence. Preset voices or your own cloned voices, in every Indian language Oogam speaks.

POST /v1/audio/speech · streamReference

Sutra

oogam.LLM

Oogam's conversation model, with native Hindi, Hinglish and Indian-language replies — through LiveKit's OpenAI-compatible LLM.

POST /v1/chat/completionsReference
How it works

Running in three steps.

01

Install

Add livekit-plugins-oogam next to LiveKit Agents. Python 3.10 or later, LiveKit Agents 1.8.

02

Add your key

Create an API key with the NAAD and SUTRA products on the Oogam Platform and set OOGAM_API_KEY.

03

Run your agent

Run python agent.py console and talk — in Hindi or English. Use dev to join a LiveKit room.

Quick start

A complete Hindi voice agent.

Oogam speech to text, Sutra and Oogam text to speech, with Silero VAD deciding when the caller has finished. Save it as agent.py and put OOGAM_API_KEY=sk-setu-… in a .env file next to it.

bash
pip install "livekit-agents[silero]~=1.8" livekit-plugins-oogam python-dotenv
python
from dotenv import load_dotenv

from livekit import agents
from livekit.agents import Agent, AgentServer, AgentSession
from livekit.plugins import oogam, silero

load_dotenv()


class Assistant(Agent):
    def __init__(self) -> None:
        super().__init__(
            instructions="You are a helpful assistant on a phone call. "
            "Reply in one or two short spoken sentences, in the caller's language.",
        )

    async def on_enter(self) -> None:
        self.session.generate_reply(instructions="Greet the caller in Hindi.")


server = AgentServer()


@server.rtc_session()
async def entrypoint(ctx: agents.JobContext):
    tts = oogam.TTS(voice="aditi", language="hi")
    session = AgentSession(
        stt=oogam.STT(language="hi", auto_switch=True, candidates=["hi", "en-IN"]),
        llm=oogam.LLM(),  # Sutra
        tts=tts,
        vad=silero.VAD.load(),
    )

    @session.on("user_input_transcribed")
    def follow_caller_language(ev):
        # Speak in the language the caller was last heard in.
        if ev.is_final and ev.language:
            tts.update_options(language=ev.language)

    await session.start(room=ctx.room, agent=Assistant())


if __name__ == "__main__":
    agents.cli.run_app(server)
bash
python agent.py console   # talk to it in your terminal
python agent.py dev       # or join a LiveKit room
Turn-taking

LiveKit decides. Oogam keeps up.

The STT plugin supports LiveKit's manual flush. When LiveKit decides the caller has stopped — Silero's end of speech, or your turn detector's verdict — it flushes the stream and the plugin sends Oogam finalize, so the turn closes at that moment. On live Hindi and Indian-English calls the transcript was ready 0.6–0.75 s after the caller stopped, and the agent started speaking 1.6–2.0 s after, Sutra and speech included.

To let Oogam end each turn by itself instead, set turn_handling={"turn_detection": "stt", "endpointing": {"min_delay": 0.07}} — the lower delay stops LiveKit waiting another half second after every transcript.

Best practices

Get the most out of it.

Keep a VAD

Pass vad=silero.VAD.load(). LiveKit flushes the STT when Silero hears the caller stop, and the plugin closes the turn on Oogam at that moment. The VAD also lets callers interrupt the agent — after the first 3 seconds of a call, while LiveKit's echo cancellation warms up (aec_warmup_duration).

Start in the caller's language

Oogam does not detect the language at the start of a call, so a session starts in the language you give it — hi, gu, ta, en-IN or a BCP-47 tag like hi-IN. If you know it from the number dialled, a menu choice or your own records, start there, and set the same language on the TTS.

Follow a switch, and move the voice

With auto_switch the session follows a caller who changes language from their second sentence — most reliably from Hindi to English today. Each transcript carries its language; move the TTS with tts.update_options(language=…), as the quick start does. A reply in another script is spoken in that script's language, so the agent is never left silent.

Size your limits

Calls are admitted per key, per account and per language, and each Sutra reply and spoken sentence counts against your key's requests-per-minute limit — about 15–25 per live call. A 429 or a 4429 close is retried by LiveKit with backoff; a bad key or an empty wallet is not retried.

Which language switches are detected today is listed under Changing language mid-call. For a pair that is not, start the session in the caller's language or switch it yourself with stt.update_options(language=…).

Sutra does not support function calling yet. If your agent uses tools, keep Oogam for speech to text and text to speech and pair them with an LLM that supports tools.

Use cases

Where teams use it.

IVR and contact centres

Answer, route and resolve calls in the caller's own language, over LiveKit SIP.

Multilingual support agents

Start each call in the caller's language — Hindi, Gujarati, Tamil, English and more — and follow a caller who moves from Hindi to English.

Outbound calling

Lead qualification, renewals and feedback calls at scale.

Resources

Everything you need.

Package
livekit-plugins-oogam · open source (MIT) · Python 3.10+ · maintained by DWET Labs
Streaming STT reference
Protocol, finalize, language switching, errors
Text to speech reference
Streaming speech · Voices
Sutra reference
Chat completions
Languages
All language codes
LiveKit
docs.livekit.io/agents

Billed from your prepaid wallet: speech to text per second the session is open, text to speech per character, Sutra per token.

Ready to build?

Create a key, install the plugin, and talk to your agent in minutes.