cekura 1.0.0rc1

Testing and Observability SDK for AI Agents

Cekura Python SDK

Testing and Observability for AI Voice Agents. Launch in minutes not weeks by ensuring your agents deliver a seamless experience in every conversational scenario.

Installation

pip install cekura

LiveKit Integration

Setup

1. Create a new agent in the Cekura dashboard
2. Select LiveKit as the provider in agent settings
3. Enable tracing for your agent
4. Copy your API key and Agent ID

Quick Start

import os
from cekura.livekit import LiveKitTracer

# Initialize Cekura tracer
cekura = LiveKitTracer(
    api_key=os.getenv("CEKURA_API_KEY"),
    agent_id=123
)

@server.rtc_session(agent_name="my_agent")
async def entrypoint(ctx: agents.JobContext):
    # Create your LiveKit session
    session = agents.AgentSession(...)

    # Start Cekura session tracking
    session_id = cekura.start_session(session)

    # Add shutdown callback to export Cekura data
    async def cekura_shutdown():
        await cekura.export(session_id, ctx)

    ctx.add_shutdown_callback(cekura_shutdown)

    # Start your session
    await session.start(room=ctx.room, agent=YourAssistant())

Next Steps

After integrating the SDK, you can:

• Run tests from the Cekura platform
• Monitor real-time conversations
• Analyze performance metrics
• Review conversation transcripts and tool calls

Features

• Automatic Metrics Collection: Captures STT, LLM, TTS, and End-of-Utterance metrics
• Conversation Tracking: Records complete chat history with tool calls
• Session Reports: Generates comprehensive session reports from LiveKit
• Easy Integration: Simple API with minimal setup required

Configuration

Parameters

• api_key (str, required): Your Cekura API key
• agent_id (int, required): Unique identifier for your agent from Cekura dashboard
• host (str, optional): API host URL (default: https://api.cekura.ai)
• enabled (bool, optional): Enable/disable tracing. Defaults to CEKURA_TRACING_ENABLED env var ("true" if not set)

Data Collected

The SDK automatically collects:

• Session start/end timestamps
• Complete conversation transcripts
• Tool/function calls and outputs
• Performance metrics:
  • Speech-to-Text (STT) latency and duration
  • LLM token usage and timing
  • Text-to-Speech (TTS) generation time
  • End-of-Utterance detection timing
• Room and job information
• Custom metadata

Requirements

• Python 3.9+
• aiohttp>=3.9.0

License

MIT License - see LICENSE file for details.

Links

• Homepage
• Documentation
• Dashboard

Support

For support, email [e-mail]