Document new LLM vendors and tighten onboarding docs.

Add Groq, Vertex AI, Bedrock, Dify, and Custom LLM to vendor references, simplify README and index navigation, and align quick-start and terminology with Agora-managed model language.

Author: digitallysavvy

README.md

@@ -7,38 +7,19 @@ The Agora Conversational AI SDK provides convenient access to the Agora Conversa
 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.
 
+## Install
 
-## Table of Contents
-
-- [Requirements](#requirements)
-- [Installation](#installation)
-- [Quick Start](#quick-start)
-- [BYOK](#byok)
-- [MLLM (Realtime / Multimodal)](#mllm-realtime--multimodal)
-- [Documentation](#documentation)
-- [Reference](#reference)
-- [Exception Handling](#exception-handling)
-- [Pagination](#pagination)
-- [Advanced](#advanced)
-  - [Access Raw Response Data](#access-raw-response-data)
-  - [Retries](#retries)
-  - [Timeouts](#timeouts)
-  - [Custom Client](#custom-client)
-- [Contributing](#contributing)
+```sh
+pip install agora-agents
+```
 
 ## Requirements
 
 - Python 3.8+
 
-## Installation
-
-```sh
-pip install agora-agents
-```
-
 ## Quick Start
 
-The recommended onboarding path is a server-side builder flow: define the agent once, configure vendors on the builder, and let AgentKit infer Agora-managed configuration when credentials are omitted.
+Start with the `Agent` builder: create a client with app credentials, choose your ASR, LLM, and TTS providers, then start a session. Omit vendor API keys for supported Agora-managed models, or provide keys when you want BYOK.
 
 ```python
 import os
@@ -146,11 +127,11 @@ def start_conversation() -> str:
 
 ### 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 inspects the builder-provided vendor configs and infers Agora-managed configuration for supported models, so you do not pass vendor API keys in this flow.
+`Agora` generates the required ConvoAI REST auth and RTC join tokens automatically when you provide `app_id` and `app_certificate`. For supported Agora-managed models, leave vendor API keys unset; provide keys when you want BYOK.
 
-### BYOK version of the same builder flow
+### BYOK version
 
-Use the same `Agent` builder shape, but provide credentials explicitly when you want vendor-managed billing and routing instead of Agora-managed presets.
+Use the same `Agent` builder shape, but provide credentials explicitly when you want vendor-managed billing and routing instead of Agora-managed models.
 
 ```python
 agent = Agent(
@@ -232,7 +213,7 @@ except ApiError as e:
 
 ## Pagination
 
-Paginated requests will return a `SyncPager` or `AsyncPager`, which can be used as generators for the underlying object.
+Paginated requests will return a `SyncPager` or `AsyncPager`, which can be used as generators for the returned object.
 
 ```python
 from agora_agent import Agora, Area
@@ -276,15 +257,15 @@ client = Agora(
 )
 response = client.agents.with_raw_response.start(...)
 print(response.headers)  # access the response headers
-print(response.data)  # access the underlying object
+print(response.data)  # access the returned 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)
+    print(item)  # access the returned 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)
+        print(item)  # access the returned object(s)
 ```
 
 ### Retries

docs/concepts/session.md

@@ -145,13 +145,13 @@ session.off('started', on_started)
 | `session.app_id` | `str` | The Agora App ID |
 | `session.raw` | `AgentsClient` | Direct access to the Fern-generated agents client |
 
-## `session.raw` — Escape Hatch
+## Direct API access with `session.raw`
 
-If the agentkit does not yet expose a method for a new API endpoint, use `session.raw` to access the underlying Fern-generated `AgentsClient` (sync) or `AsyncAgentsClient` (async) directly:
+If AgentKit does not yet expose a method for a new API endpoint, use `session.raw` to access the generated `AgentsClient` (sync) or `AsyncAgentsClient` (async) directly:
 
 <!-- snippet: fragment -->
 ```python
-# Access any Fern-generated method
+# Access any generated REST method
 response = session.raw.list(session.app_id)
 ```
 

docs/concepts/vendors.md

@@ -25,6 +25,11 @@ Used with `agent.with_llm()` for the cascading flow (ASR → LLM → TTS).
 | `AzureOpenAI` | Azure OpenAI | `api_key`, `endpoint`, `deployment_name` |
 | `Anthropic` | Anthropic | `api_key` |
 | `Gemini` | Google Gemini | `api_key` |
+| `Groq` | Groq | `api_key` |
+| `VertexAILLM` | Google Vertex AI | `api_key`, `project_id`, `location` |
+| `AmazonBedrock` | Amazon Bedrock | `api_key`, `url`, `model` |
+| `Dify` | Dify | `api_key`, `url` |
+| `CustomLLM` | OpenAI-compatible LLM | `api_key`, `base_url`, `model` |
 
 <!-- snippet: executable -->
 ```python

docs/getting-started/quick-start.md

@@ -6,7 +6,7 @@ description: Build and run your first Agora Conversational AI agent in Python wi
 
 # Quick Start
 
-This guide uses the recommended onboarding path:
+This guide starts with the standard AgentKit path:
 
 - `app_id`, `app_certificate`, and `area` on `Agora` or `AsyncAgora`
 - the `Agent` builder with `.with_stt()`, `.with_llm()`, and `.with_tts()`
@@ -61,7 +61,7 @@ if __name__ == "__main__":
 
 1. `Agora` runs in app-credentials mode when you pass `app_id` and `app_certificate` only.
 2. `Agent` holds reusable behavior such as instructions, greeting, and history settings.
-3. Vendor classes on the builder select the ASR, LLM, and TTS stack. AgentKit infers Agora-managed configuration when credentials are omitted for supported models.
+3. Vendor classes on the builder select the ASR, LLM, and TTS stack. Leave vendor credentials unset for supported Agora-managed models, or provide keys when you want BYOK.
 4. `session.start()` generates the required auth tokens and returns the unique agent session ID.
 
 ## Async applications
@@ -70,7 +70,7 @@ For `asyncio` services, switch to `AsyncAgora` and `await` the session methods.
 
 ## When to use BYOK instead
 
-Use the builder without vendor API keys when you want the fastest path with Agora-managed models.
+Use the builder without vendor API keys when you are using supported Agora-managed models.
 
 Use BYOK when you need to:
 

docs/guides/byok.md

@@ -81,5 +81,5 @@ if __name__ == "__main__":
 
 ## Builder-managed vs BYOK
 
-- Builder without vendor keys: fastest path with Agora-managed models
+- Builder without vendor keys: supported Agora-managed models
 - BYOK: your keys and full vendor control

docs/guides/regional-routing.md

@@ -108,7 +108,7 @@ print(client.get_current_url())
 
 ## Accessing the Pool Directly
 
-For advanced use, access the underlying `Pool` object:
+For advanced use, access the `Pool` object:
 
 ```python
 pool = client.pool
@@ -127,4 +127,4 @@ area = pool.get_area()
 | `next_region()` | sync | sync | Cycle to next region prefix |
 | `select_best_domain()` | sync | **`async` (requires `await`)** | DNS-based domain selection |
 | `get_current_url()` | sync | sync | Get the current base URL |
-| `pool` (property) | sync | sync | Access the underlying `Pool` object |
+| `pool` (property) | sync | sync | Access the `Pool` object |

docs/index.md

@@ -19,29 +19,29 @@ The Agora Conversational AI Python SDK lets you build voice-powered AI agents on
 
 **MLLM flow** uses a multimodal model such as OpenAI Realtime, Gemini Live, Vertex AI, or xAI Grok for end-to-end audio.
 
-## Choose a starting point
+## Start here
 
-- Use [Quick Start](./getting-started/quick-start.md) if you want the recommended builder-based path with app credentials.
-- Use [MLLM Flow](./guides/mllm-flow.md) if you want realtime end-to-end audio with OpenAI Realtime, Gemini Live, Vertex AI, or xAI Grok.
-- Use [Cascading Flow](./guides/cascading-flow.md) if you want separate ASR, LLM, and TTS vendors.
+- Start with [Quick Start](./getting-started/quick-start.md). It shows the baseline app-credentials setup and starts a cascading ASR -> LLM -> TTS agent.
+- Use [MLLM Flow](./guides/mllm-flow.md) when your agent uses one realtime multimodal model, such as OpenAI Realtime, Gemini Live, Vertex AI, or xAI Grok.
+- Use [Cascading Flow](./guides/cascading-flow.md) for more examples of the default ASR -> LLM -> TTS flow, including provider-specific configuration.
 
-## SDK layers
+## How the SDK is organized
 
 | Layer | What it does | When to use |
 |---|---|---|
 | **AgentKit** (`Agent`, `AgentSession`, vendor classes) | High-level builder pattern, lifecycle, typed vendors | Most use cases |
-| **Fern-generated core** (`client.agents`, `client.telephony`) | Direct REST client mapping every API endpoint | Advanced use cases |
+| **Generated REST clients** (`client.agents`, `client.telephony`) | Typed access to REST APIs not covered by AgentKit | Advanced use cases |
 
 ## Documentation
 
 | Section | What you will learn |
 |---|---|
 | [Installation](./getting-started/installation.md) | Install the SDK and prerequisites |
 | [Authentication](./getting-started/authentication.md) | App credentials and other auth modes |
-| [Quick Start](./getting-started/quick-start.md) | Recommended builder-based onboarding flow |
+| [Quick Start](./getting-started/quick-start.md) | App credentials and AgentKit |
 | [Agent Builder Features](./guides/agent-builder-features.md) | Turn detection, SAL, filler words, and advanced agent options |
 | [BYOK](./guides/byok.md) | Bring your own vendor credentials and config |
-| [Architecture](./concepts/architecture.md) | Understand the SDK layers and client types |
+| [Architecture](./concepts/architecture.md) | SDK structure and generated REST clients |
 | [Agent](./concepts/agent.md) | Configure agents with the fluent builder |
 | [AgentSession](./concepts/session.md) | Manage the agent lifecycle |
 | [Vendors](./concepts/vendors.md) | Browse all LLM, TTS, STT, MLLM, and Avatar providers |
@@ -52,7 +52,7 @@ The Agora Conversational AI Python SDK lets you build voice-powered AI agents on
 | [Error Handling](./guides/error-handling.md) | Handle API errors with ApiError |
 | [Pagination](./guides/pagination.md) | Iterate over paginated list endpoints |
 | [Advanced](./guides/advanced.md) | Raw response, retries, timeouts, custom httpx client |
-| [Low-Level API](./guides/low-level-api.md) | Direct `client.agents.start()` usage |
+| [Low-Level API](./guides/low-level-api.md) | Generated REST APIs |
 | [Client Reference](./reference/client.md) | Full `Agora` / `AsyncAgora` API |
 | [Agent Reference](./reference/agent.md) | Full `Agent` builder API |
 | [Session Reference](./reference/session.md) | Full `AgentSession` / `AsyncAgentSession` API |

docs/reference/client.md

@@ -100,7 +100,7 @@ url = client.get_current_url()
 
 ### `pool` (property)
 
-Access the underlying `Pool` object for advanced domain management.
+Access the `Pool` object for advanced domain management.
 
 <!-- snippet: fragment -->
 ```python

docs/reference/vendors.md

@@ -131,6 +131,18 @@ from agora_agent import Gemini
 llm = Gemini(api_key='your-google-key', model='gemini-2.0-flash-exp')
 ```
 
+### Other LLM vendors
+
+The SDK also includes named helpers for the remaining Agora-supported LLM providers. These helpers choose the correct request format internally.
+
+| Class | Provider | Key parameters |
+|---|---|---|
+| `Groq` | Groq | `api_key`, `model`, `base_url?` |
+| `VertexAILLM` | Google Vertex AI | `api_key`, `model`, `project_id`, `location`, `url?` |
+| `AmazonBedrock` | Amazon Bedrock | `api_key`, `url`, `model` |
+| `Dify` | Dify | `api_key`, `url`, `user?`, `conversation_id?` |
+| `CustomLLM` | OpenAI-compatible LLM | `api_key`, `model`, `base_url` |
+
 ---
 
 ## TTS Vendors