Integrations

Build with the tools you already use.

Oogam speech to text, text to speech and Sutra, ready to drop into the voice-agent frameworks teams already build on. Each integration is an open-source Python package, tested on live calls in Hindi, Indian English, Tamil and Gujarati.

Pipecat

Agent framework

An open-source Python framework that runs your voice agent as a pipeline: audio in, speech to text, LLM, text to speech, audio out.

Package
pipecat-oogam
Includes
OogamSTTServiceOogamTTSServiceOogamLLMService
Needs
Python 3.11+ · Pipecat 1.16+
bash
pip install \
  "pipecat-ai[runner,webrtc,silero]" \
  pipecat-oogam

LiveKit

Voice infrastructure

Real-time audio rooms, telephony and SIP, with LiveKit Agents to run your voice agent inside a room or on a phone call.

Package
livekit-plugins-oogam
Includes
oogam.STToogam.TTSoogam.LLM
Needs
Python 3.10+ · LiveKit Agents 1.8
bash
pip install \
  "livekit-agents[silero]~=1.8" \
  livekit-plugins-oogam
Which one?

Both give you the same Oogam models.

The speech, languages, voices and Sutra are the same in either package. The difference is how your agent runs and how callers reach it.

Choose Pipecat if

  • you want one Python process you run anywhere — your laptop, a VM or a container
  • you want to see and change every step of the pipeline
  • your callers reach the agent from a browser, or from a phone provider that streams audio over WebSockets

Choose LiveKit if

  • you want rooms with several participants, recording and real-time media handled for you
  • you want phone numbers and SIP trunks through LiveKit SIP
  • you want agent workers that LiveKit dispatches and scales, on LiveKit Cloud or your own server
Something else?

Call the APIs directly.

Using another framework, or none? The integrations are thin layers over these public APIs — any language that speaks WebSockets and HTTP can use them.

Ready to build?

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