> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dappier.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Google ADK

## Overview

The integration of **Google ADK** and **Dappier SDK** empowers developers to rapidly build **multi-agent AI applications** that combine Google’s modular agent orchestration with Dappier’s real-time, rights-cleared data sources. By pairing Gemini-powered agents with Dappier’s live insights, developers can create **smart assistants**, **dynamic analytics tools**, and **automated content recommendation systems** that operate with **accuracy, adaptability, and speed**.

## Watch the Video

If you prefer a visual walkthrough, check out the accompanying video below:

<iframe width="560" height="315" src="https://www.youtube.com/embed/4jza4LSdgLY?si=X-QKS8HD7RxIdKRV" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen />

## Starter App: Google ADK + Dappier

To explore the capabilities of integrating **Google ADK** with **Dappier**, check out this ready-to-run starter project:

👉 [Google ADK + Dappier Starter App on Replit](https://replit.com/@dappier/Google-ADK-Dappier-Starter-App?v=1)

This application demonstrates how to:

* Utilize Google ADK’s `Agent` class to build AI agents powered by Gemini models.
* Integrate Dappier’s real-time search and AI recommendation tools to fetch up-to-date information across various domains, including finance, sports, lifestyle, and pet care.
* Configure environment variables for seamless authentication and deployment.

Feel free to explore and customize this template to suit your specific use cases.

## Google ADK

**Google’s Agent Development Kit (ADK)** is an open-source Python framework designed to help developers build, evaluate, and deploy **modular, multi-agent AI systems**. It is optimized for Google’s **Gemini models**, but remains model-agnostic and flexible across a wide range of LLMs and environments.

### Features of Google ADK:

* **Agent Composition** – Build agents with modular tools and behaviors, including Sequential, Parallel, Loop, and LLM-directed control flows.
* **Tool Integration** – Equip agents with powerful tools for tasks like web search, code execution, or third-party APIs.
* **Context Handling** – Maintain session state, memory, and context for dynamic agent interactions.
* **Evaluation Framework** – Assess agent accuracy and trace step-by-step reasoning during testing.
* **Flexible Deployment** – Run locally, on Google Cloud (e.g., Cloud Run, Vertex AI), or in serverless environments like Replit.

> Learn more at [google.github.io/adk-docs](https://google.github.io/adk-docs/)

## Dappier

**Dappier SDK** is a real-time AI data platform that connects AI agents to **verified, rights-cleared content and live information** from the web. It empowers developers to build **hallucination-free**, trustworthy, and context-aware AI applications.

### Key Capabilities of Dappier:

* **Real-Time Web Search** – Retrieve up-to-the-minute data across categories like news, finance, weather, travel, events, and more.
* **AI Recommendations** – Deliver personalized content from niche domains such as sports, lifestyle, pet care, and sustainability.
* **Data Integrity** – Powered by rights-cleared sources to ensure factual correctness and legal safety in generative AI outputs.
* **Marketplace Models** – Access dozens of plug-and-play data models for use in AI workflows.

> **Explore live data models at** [marketplace.dappier.com](https://marketplace.dappier.com)

## Getting Started: Build the Google ADK + Dappier Starter App

Follow these simple steps to run the starter app using **Google ADK + Dappier SDK** — no installation required.

### 🔹 Step 1: Fork the Starter App on Replit

1. Open the project in your browser:\
   👉 [Google ADK + Dappier Starter App](https://replit.com/@dappier/Google-ADK-Dappier-Starter-App?v=1)

2. Click **"Fork"** at the top-right to create your own copy.

3. Replit will automatically set up the project environment and open the code editor.

Once forked, you're ready to configure your API keys.

Here’s **Step 2**, guiding users to set up their API keys securely in Replit:

### 🔹 Step 2: Add API Keys via Secrets (Replit)

To authenticate with Google and Dappier services, you’ll need to set up three environment variables in Replit:

1. In your Replit workspace, click the **🔐 "Secrets" (lock icon)** in the left sidebar.
2. Add the following environment variables:

| Name                        | Description                                                                                     |
| --------------------------- | ----------------------------------------------------------------------------------------------- |
| `GOOGLE_API_KEY`            | Your Google API Key for Gemini models. [Get one here](https://makersuite.google.com/app/apikey) |
| `GOOGLE_GENAI_USE_VERTEXAI` | Set this to `"false"` unless you're using Vertex AI.                                            |
| `DAPPIER_API_KEY`           | Your Dappier API Key. [Generate it here](https://platform.dappier.com/profile/api-keys)         |

3. Once added, Replit will securely inject them into your code environment.

> ⚠️ Never hardcode API keys in your code. Use environment variables for security.

Here’s **Step 3**, helping users understand how the agent and Dappier-powered tools are defined and connected:

### 🔹 Step 3: Understand the Agent and Tool Functions

At the core of this app is a **Google ADK `Agent`**, powered by **Gemini** and enhanced using **Dappier tools**.

#### 🧠 Root Agent Setup

```python theme={null}
from google.adk.agents import Agent

root_agent = Agent(
    name="dappier_tools_agent",
    model="gemini-2.0-flash",
    description="Agent to answer questions using Dappier tools.",
    instruction="You are a helpful assistant that uses Dappier tools to fetch live data.",
    tools=[
        real_time_web_search,
        stock_market_data_search,
        get_sports_news,
        get_lifestyle_news,
        get_iheartdogs_content,
        get_iheartcats_content,
        get_greenmonster_guides
    ]
)
```

This `Agent` is equipped with multiple tool functions that fetch **live, trusted data** using Dappier.

#### 🔧 Sample Tool Function

```python theme={null}
def real_time_web_search(query: str) -> str:
    return client.search_real_time_data_string(
        query=query,
        ai_model_id="am_01j06ytn18ejftedz6dyhz2b15"
    )
```

Each tool is mapped to a specific **Dappier AI model**, like real-time search, stock updates, lifestyle news, and more.

In the next step, we’ll run the agent and interact with it.

### 🔹 Step 4: Run and Interact with the Agent

Since the Replit "Run" button won't launch the ADK server, you'll need to start it manually using the terminal.

#### ▶ How to Run

1. Click the **"Run"** button at the top of your Replit workspace.

2. Replit will display a live URL such as:
   `https://<your-replit-username>.<repl-name>.repl.co`
   Click on this URL to open the **interactive agent UI** in a new browser tab.

#### 💬 Sample Prompts to Try

Once the interface is running, try asking your agent:

* `"What's the latest update on Tesla stock?"`
* `"Give me trending dog health tips"`
* `"Show me eco-friendly lifestyle news"`
* `"What's happening in sports today?"`

The agent will intelligently route your query to the appropriate **Dappier-powered tool**, fetch real-time data, and respond using **Gemini-generated answers**.

Your custom AI assistant is now live and running on the web! ✅

## Why Integrate Google ADK and Dappier?

Combining **Google ADK** with **Dappier SDK** brings together the strengths of **modular AI agent orchestration** and **real-time, rights-cleared data access**. This integration unlocks the ability to build **dynamic, context-aware agents** that can reason, route tasks, and respond with up-to-the-minute information.

### 🔍 Key Benefits

* **Real-Time Decision Making**\
  Agents can base responses on the latest stock prices, breaking news, or trending lifestyle content — not static memory.

* **Reliable AI Outputs**\
  Reduce hallucinations by fetching data from trusted sources using Dappier’s verified pipelines.

* **Multi-Agent Workflows**\
  Use Google ADK’s support for `Sequential`, `Parallel`, or LLM-directed flows to design complex, modular agents.

* **Diverse Use Cases**\
  Easily support finance, travel, sports, sustainability, and more — with plug-and-play data models from Dappier.

### 💡 Example Scenarios

1. **Stock Market Assistant**\
   Track tickers like TSLA or AAPL in real time and generate Gemini-based insights.

2. **Pet Care Advisor**\
   Recommend dog or cat health content from expert-curated sources like iHeartDogs or iHeartCats.

3. **Lifestyle Curator**\
   Pull trending lifestyle updates or conscious living tips from GreenMonster and other top publications.

4. **Smart News Aggregator**\
   Build a conversational news assistant that fetches articles based on keywords and personal interest.

Together, Google ADK and Dappier empower you to build AI agents that are not only smart — but **informed, safe, and ready for production**.

## Conclusion

The integration of **Google ADK** with **Dappier SDK** gives developers a powerful foundation to build **real-time, intelligent AI agents** that are both modular and grounded in trusted data.

Whether you're building a **stock advisor**, **news assistant**, **travel planner**, or **lifestyle recommender**, this stack ensures your agents are:

* Powered by **Gemini models** via Google ADK
* Informed by **live, verified data** via Dappier
* Deployable anywhere — including **Replit**, **Cloud Run**, and beyond

> Try the starter template today:\
> 👉 [Google ADK + Dappier Starter App](https://replit.com/@dappier/Google-ADK-Dappier-Starter-App?v=1)

Ready to build the next generation of real-world AI tools?\
This is your launchpad.