Agoraio Python Library

The Agora Conversational AI SDK provides convenient access to the Agora Conversational AI APIs, enabling you to build voice-powered AI agents with support for both cascading flows (ASR -> LLM -> TTS) and multimodal flows (MLLM) for real-time audio processing.

Table of Contents

• Requirements
• Installation
• Quick Start
• Byok
• Mllm Realtime Multimodal
• Documentation
• Reference
• Mllm Flow Multimodal
• Mllm Flow Multimodal
• Usage
• Async Client
• Exception Handling
• Pagination
• Advanced
  • Access Raw Response Data
  • Retries
  • Timeouts
  • Custom Client
• Contributing

Requirements

• Python 3.8+

Installation

pip install chenyuguo-agora-agents-exp

Quick Start

The recommended onboarding path is a server-side builder flow: define the agent once, configure preset-backed providers in the builder, and let AgentKit infer the reseller preset values when the session starts.

import os
import time

from agora_agent import Agora, Area
from agora_agent.agentkit import (
    Agent,
    DataChannel,
    DeepgramSTT,
    GenericAvatar,
    MiniMaxTTS,
    OpenAI,
    XaiGrok,
    expires_in_hours,
)

AGENT_PROMPT = (
    "You are a concise, technically credible voice assistant. "
    "Keep replies short unless the user asks for detail."
)

GREETING = "Hi there! I am your Agora voice assistant. How can I help?"


def start_conversation() -> str:
    app_id = os.environ["AGORA_APP_ID"]
    app_certificate = os.environ["AGORA_APP_CERTIFICATE"]

    client = Agora(
        area=Area.US,
        app_id=app_id,
        app_certificate=app_certificate,
    )

    agent = Agent(
        name=f"conversation-{int(time.time())}",
        instructions=AGENT_PROMPT,
        greeting=GREETING,
        failure_message="Please wait a moment.",
        max_history=50,
        turn_detection={
            "config": {
                "speech_threshold": 0.5,
                "start_of_speech": {
                    "mode": "vad",
                    "vad_config": {
                        "interrupt_duration_ms": 160,
                        "prefix_padding_ms": 300,
                    },
                },
                "end_of_speech": {
                    "mode": "vad",
                    "vad_config": {
                        "silence_duration_ms": 480,
                    },
                },
            },
        },
        advanced_features={
            "enable_rtm": True,
            "enable_tools": True,
        },
        parameters={
            "data_channel": DataChannel.RTM,
            "enable_error_message": True,
        },
    ).with_stt(
        DeepgramSTT(
            model="nova-3",
            language="en",
        )
    ).with_llm(
        OpenAI(
            model="gpt-4o-mini",
            greeting_message=GREETING,
            failure_message="Please wait a moment.",
            max_history=15,
            params={
                "max_tokens": 1024,
                "temperature": 0.7,
                "top_p": 0.95,
            },
        )
    ).with_tts(
        MiniMaxTTS(
            model="speech_2_6_turbo",
            voice_id="English_captivating_female1",
        )
    )

    session = agent.create_session(
        client,
        channel=f"demo-channel-{int(time.time())}",
        agent_uid="123456",
        remote_uids=["*"],
        idle_timeout=30,
        expires_in=expires_in_hours(1),
        debug=False,
    )

    return session.start()

Why no token or vendor key in the example?

Agora generates the required ConvoAI REST auth and RTC join tokens automatically when you provide app_id and app_certificate. AgentKit then inspects the builder-provided vendor configs and infers the matching supported preset values for reseller-backed models, so you do not pass vendor API keys in this flow.

BYOK version of the same builder flow

Use the same Agent builder shape, but provide credentials explicitly when you want vendor-managed billing and routing instead of Agora-managed presets.

agent = Agent(
    instructions=AGENT_PROMPT,
    greeting=GREETING,
).with_stt(
    DeepgramSTT(
        api_key=os.environ["DEEPGRAM_API_KEY"],
        model="nova-3",
        language="en",
    )
).with_llm(
    OpenAI(
        api_key=os.environ["OPENAI_API_KEY"],
        model="gpt-4o-mini",
        max_tokens=1024,
        temperature=0.7,
        top_p=0.95,
    )
).with_tts(
    MiniMaxTTS(
        key=os.environ["MINIMAX_API_KEY"],
        group_id=os.environ["MINIMAX_GROUP_ID"],
        model="speech_2_6_turbo",
        voice_id="English_captivating_female1",
        url="wss://api-uw.minimax.io/ws/v1/t2a_v2",
    )
)

BYOK

If you want to bring your own vendor credentials instead of using Agora-managed presets, use the BYOK guide:

• BYOK Guide

MLLM (Realtime / Multimodal)

Use with_mllm() for OpenAI Realtime, Gemini Live, Vertex AI, or xAI Grok. No STT, LLM, or TTS vendor is needed when MLLM mode is enabled.

from agora_agent.agentkit import Agent, OpenAIRealtime

agent = Agent(name="realtime-assistant").with_mllm(
    OpenAIRealtime(
        api_key=os.environ["OPENAI_API_KEY"],
        model="gpt-4o-realtime-preview",
        greeting_message="Hello! Ready to chat.",
    )
)

See the MLLM Flow guide for full examples with Gemini Live and Vertex AI.

Documentation

API reference documentation is available here.

Reference

A full reference for this library is available here.

Package Rename Compatibility

The published package name is now chenyuguo-agora-agents-exp, while the Python import path remains agora_agent for compatibility. The legacy PyPI distribution name chenyuguo-agora-agent-server-sdk-exp is maintained as a compatibility package in compat/agora-agent-server-sdk, and the tag-based release workflow publishes both distributions together.

MLLM Flow (Multimodal)

For real-time audio processing using OpenAI Realtime, Gemini Live, Vertex AI, or xAI Grok, use the MLLM (Multimodal Large Language Model) flow instead of the cascading ASR -> LLM -> TTS flow. MLLM mode does not require separate TTS, STT, or LLM vendors. See the MLLM Overview for more details.

from agora_agent import Agora
from agora_agent.agents import (
    StartAgentsRequestProperties,
    StartAgentsRequestPropertiesMllm,
    StartAgentsRequestPropertiesMllmVendor,
    StartAgentsRequestPropertiesTurnDetection,
    StartAgentsRequestPropertiesTurnDetectionType,
)

client = Agora(
    customer_id="YOUR_CUSTOMER_ID",
    customer_secret="YOUR_CUSTOMER_SECRET",
)

client.agents.start(
    appid="your_app_id",
    name="mllm_agent",
    properties=StartAgentsRequestProperties(
        channel="channel_name",
        token="your_token",
        agent_rtc_uid="1001",
        remote_rtc_uids=["1002"],
        idle_timeout=120,
        mllm=StartAgentsRequestPropertiesMllm(
            enable=True,
            url="wss://api.openai.com/v1/realtime",
            api_key="<your_openai_api_key>",
            vendor=StartAgentsRequestPropertiesMllmVendor.OPENAI,
            params={
                "model": "gpt-4o-realtime-preview",
                "voice": "alloy",
            },
            input_modalities=["audio"],
            output_modalities=["text", "audio"],
            greeting_message="Hello! I'm ready to chat in real-time.",
        ),
        turn_detection=StartAgentsRequestPropertiesTurnDetection(
            type=StartAgentsRequestPropertiesTurnDetectionType.SERVER_VAD,
            threshold=0.5,
            silence_duration_ms=500,
        ),
    ),
)

MLLM Flow (Multimodal)

For real-time audio processing using OpenAI Realtime, Gemini Live, Vertex AI, or xAI Grok, use the MLLM (Multimodal Large Language Model) flow instead of the cascading ASR -> LLM -> TTS flow. MLLM mode does not require separate TTS, STT, or LLM vendors. See the MLLM Overview for more details.

from agora_agent import Agora
from agora_agent.agents import (
    StartAgentsRequestProperties,
    StartAgentsRequestPropertiesMllm,
    StartAgentsRequestPropertiesMllmVendor,
    StartAgentsRequestPropertiesTurnDetection,
    StartAgentsRequestPropertiesTurnDetectionType,
)

client = Agora(
    customer_id="YOUR_CUSTOMER_ID",
    customer_secret="YOUR_CUSTOMER_SECRET",
)

client.agents.start(
    appid="your_app_id",
    name="mllm_agent",
    properties=StartAgentsRequestProperties(
        channel="channel_name",
        token="your_token",
        agent_rtc_uid="1001",
        remote_rtc_uids=["1002"],
        idle_timeout=120,
        mllm=StartAgentsRequestPropertiesMllm(
            enable=True,
            url="wss://api.openai.com/v1/realtime",
            api_key="<your_openai_api_key>",
            vendor=StartAgentsRequestPropertiesMllmVendor.OPENAI,
            params={
                "model": "gpt-4o-realtime-preview",
                "voice": "alloy",
            },
            input_modalities=["audio"],
            output_modalities=["text", "audio"],
            greeting_message="Hello! I'm ready to chat in real-time.",
        ),
        turn_detection=StartAgentsRequestPropertiesTurnDetection(
            type=StartAgentsRequestPropertiesTurnDetectionType.SERVER_VAD,
            threshold=0.5,
            silence_duration_ms=500,
        ),
    ),
)

Usage

Instantiate and use the client with the following:

from agora_agent import Agora, MicrosoftTtsParams, Tts_Microsoft
from agora_agent.agents import (
    StartAgentsRequestProperties,
    StartAgentsRequestPropertiesAsr,
    StartAgentsRequestPropertiesLlm,
    StartAgentsRequestPropertiesTurnDetection,
    StartAgentsRequestPropertiesTurnDetectionConfig,
    StartAgentsRequestPropertiesTurnDetectionConfigEndOfSpeech,
)

client = Agora(
    authorization="YOUR_AUTHORIZATION",
    username="YOUR_USERNAME",
    password="YOUR_PASSWORD",
)
client.agents.start(
    appid="appid",
    name="unique_name",
    properties=StartAgentsRequestProperties(
        channel="channel_name",
        token="token",
        agent_rtc_uid="1001",
        remote_rtc_uids=["1002"],
        idle_timeout=120,
        asr=StartAgentsRequestPropertiesAsr(
            language="en-US",
        ),
        tts=Tts_Microsoft(
            params=MicrosoftTtsParams(
                key="key",
                region="region",
                voice_name="voice_name",
            ),
        ),
        llm=StartAgentsRequestPropertiesLlm(
            url="https://api.openai.com/v1/chat/completions",
            api_key="<your_llm_key>",
            system_messages=[
                {"role": "system", "content": "You are a helpful chatbot."}
            ],
            params={"model": "gpt-4o-mini"},
            max_history=32,
            greeting_message="Hello, how can I assist you today?",
            failure_message="Please hold on a second.",
        ),
        turn_detection=StartAgentsRequestPropertiesTurnDetection(
            config=StartAgentsRequestPropertiesTurnDetectionConfig(
                end_of_speech=StartAgentsRequestPropertiesTurnDetectionConfigEndOfSpeech(
                    mode="semantic",
                ),
            ),
        ),
    ),
)

Async Client

The SDK also exports an async client so that you can make non-blocking calls to our API. Note that if you are constructing an Async httpx client class to pass into this client, use httpx.AsyncClient() instead of httpx.Client() (e.g. for the httpx_client parameter of this client).

import asyncio

from agora_agent import AsyncAgora, MicrosoftTtsParams, Tts_Microsoft
from agora_agent.agents import (
    StartAgentsRequestProperties,
    StartAgentsRequestPropertiesAsr,
    StartAgentsRequestPropertiesLlm,
    StartAgentsRequestPropertiesTurnDetection,
    StartAgentsRequestPropertiesTurnDetectionConfig,
    StartAgentsRequestPropertiesTurnDetectionConfigEndOfSpeech,
)

client = AsyncAgora(
    authorization="YOUR_AUTHORIZATION",
    username="YOUR_USERNAME",
    password="YOUR_PASSWORD",
)


async def main() -> None:
    await client.agents.start(
        appid="appid",
        name="unique_name",
        properties=StartAgentsRequestProperties(
            channel="channel_name",
            token="token",
            agent_rtc_uid="1001",
            remote_rtc_uids=["1002"],
            idle_timeout=120,
            asr=StartAgentsRequestPropertiesAsr(
                language="en-US",
            ),
            tts=Tts_Microsoft(
                params=MicrosoftTtsParams(
                    key="key",
                    region="region",
                    voice_name="voice_name",
                ),
            ),
            llm=StartAgentsRequestPropertiesLlm(
                url="https://api.openai.com/v1/chat/completions",
                api_key="<your_llm_key>",
                system_messages=[
                    {"role": "system", "content": "You are a helpful chatbot."}
                ],
                params={"model": "gpt-4o-mini"},
                max_history=32,
                greeting_message="Hello, how can I assist you today?",
                failure_message="Please hold on a second.",
            ),
            turn_detection=StartAgentsRequestPropertiesTurnDetection(
                config=StartAgentsRequestPropertiesTurnDetectionConfig(
                    end_of_speech=StartAgentsRequestPropertiesTurnDetectionConfigEndOfSpeech(
                        mode="semantic",
                    ),
                ),
            ),
        ),
    )


asyncio.run(main())

Exception Handling

When the API returns a non-success status code (4xx or 5xx response), a subclass of the following error will be thrown.

from agora_agent.core.api_error import ApiError

try:
    client.agents.start(...)
except ApiError as e:
    print(e.status_code)
    print(e.body)

Pagination

Paginated requests will return a SyncPager or AsyncPager, which can be used as generators for the underlying object.

from agora_agent import Agora

client = Agora(
    authorization="YOUR_AUTHORIZATION",
    username="YOUR_USERNAME",
    password="YOUR_PASSWORD",
)
response = client.agents.list(
    appid="appid",
)
for item in response:
    yield item
# alternatively, you can paginate page-by-page
for page in response.iter_pages():
    yield page

# You can also iterate through pages and access the typed response per page
pager = client.agents.list(...)
for page in pager.iter_pages():
    print(page.response)  # access the typed response for each page
    for item in page:
        print(item)

Advanced

Access Raw Response Data

The SDK provides access to raw response data, including headers, through the .with_raw_response property. The .with_raw_response property returns a "raw" client that can be used to access the .headers and .data attributes.

from agora_agent import Agora

client = Agora(
    ...,
)
response = client.agents.with_raw_response.start(...)
print(response.headers)  # access the response headers
print(response.data)  # access the underlying object
pager = client.agents.list(...)
print(pager.response)  # access the typed response for the first page
for item in pager:
    print(item)  # access the underlying object(s)
for page in pager.iter_pages():
    print(page.response)  # access the typed response for each page
    for item in page:
        print(item)  # access the underlying object(s)

Retries

The SDK is instrumented with automatic retries with exponential backoff. A request will be retried as long as the request is deemed retryable and the number of retry attempts has not grown larger than the configured retry limit (default: 2).

A request is deemed retryable when any of the following HTTP status codes is returned:

• 408 (Timeout)
• 429 (Too Many Requests)
• 5XX (Internal Server Errors)

Use the max_retries request option to configure this behavior.

client.agents.start(..., request_options={
    "max_retries": 1
})

Timeouts

The SDK defaults to a 60 second timeout. You can configure this with a timeout option at the client or request level.

from agora_agent import Agora

client = Agora(
    ...,
    timeout=20.0,
)


# Override timeout for a specific method
client.agents.start(..., request_options={
    "timeout_in_seconds": 1
})

Custom Client

You can override the httpx client to customize it for your use-case. Some common use-cases include support for proxies and transports.

import httpx
from agora_agent import Agora

client = Agora(
    ...,
    httpx_client=httpx.Client(
        proxy="http://my.test.proxy.example.com",
        transport=httpx.HTTPTransport(local_address="0.0.0.0"),
    ),
)

Contributing

While we value open-source contributions to this SDK, this library is generated programmatically. Additions made directly to this library would have to be moved over to our generation code, otherwise they would be overwritten upon the next generated release. Feel free to open a PR as a proof of concept, but know that we will not be able to merge it as-is. We suggest opening an issue first to discuss with us!

On the other hand, contributions to the README are always very welcome!