Voice Agents

Agent Options

The repository contains two voice-agent implementations.

File	Stack	Best use
`agent.py`	Sarvam STT + Sarvam-M LLM + Sarvam TTS	Indian-language pipeline with separate speech components
`agent_realtime.py`	OpenAI Realtime	Low-latency all-in-one realtime conversation

Run one agent service at a time for the calling flow you want.

Sarvam Agent

Command:

python agent.py start

Pipeline:

Lead audio
  -> Sarvam STT
  -> Sarvam-M LLM
  -> Sarvam TTS
  -> LiveKit audio response

Key settings:

SARVAM_API_KEY=your_sarvam_api_key
STT_LANGUAGE=en-IN
STT_MODEL=saarika:v2.5
STT_FLUSH_SIGNAL=true
MIN_ENDPOINTING_DELAY_SECONDS=0.65

Qualification fields:

machine_interest
location
budget
first_product
units_per_day
local_demand
new_or_expand
partnership
operators_needed
own_brand
seriousness
factory_visit
video_demo

OpenAI Realtime Agent

Command:

python agent_realtime.py start

Pipeline:

Lead audio
  -> OpenAI Realtime session
  -> model handles listening, reasoning, and speaking
  -> LiveKit audio response

Key settings:

OPENAI_API_KEY=your_openai_api_key
OPENAI_REALTIME_MODEL=gpt-4o-realtime-preview
OPENAI_REALTIME_TRANSCRIBE_MODEL=whisper-1
OPENAI_REALTIME_TEMPERATURE=0.5
CALL_MAX_DURATION_SECONDS=240

Qualification fields:

product_interest: carbon cleaning, battery regeneration, car wash, franchise, or unsure
location
budget: below 10 lakhs, 10-25 lakhs, or >25 lakhs

Save Behavior

Both agents:

Build a transcript/history during the call.
Extract structured qualification from the transcript.
Remove null values.
Save the JSON to calls.qualification.
Mark the call called.
Hang up or finish the room.

If the participant disconnects first, the agent attempts a partial save. If no useful answers exist, it marks the call failed and can schedule a retry.

Language Behavior

agent_realtime.py contains detection and reply guidance for English and Indian-language user intent, including local-language and romanized-language handling. It also normalizes saved qualification values back to English-only structured data so downstream CRM payloads stay consistent.

Call Limits

OpenAI Realtime mode watches maximum call age:

CALL_MAX_DURATION_SECONDS=240

When the maximum duration is reached, the agent saves available qualification data and hangs up.

Getting Started

Architecture

Integrations

Operations

Agent Options

Sarvam Agent

OpenAI Realtime Agent

Save Behavior

Language Behavior

Call Limits

​Agent Options

​Sarvam Agent

​OpenAI Realtime Agent

​Save Behavior

​Language Behavior

​Call Limits

Agent Options

Sarvam Agent

OpenAI Realtime Agent

Save Behavior

Language Behavior

Call Limits