mirror of
https://github.com/google/adk-docs.git
synced 2026-09-14 16:16:59 +08:00
e25527b1dd
* Create enterprise-web-search.md Rendered page: Logo's PR: #2057 Agent's PR: #1121 and #1122 Original Issue: #1113 -- This PR adds the official documentation for the new EnterpriseWebSearchTool. Key additions: - Overview & Use Cases: Explains enterprise compliance, controlled web access, and regulated workflows. - Code Examples checked against repositories: Provides initialization snippets for both Python (google-adk) and TypeScript (@google/adk). * Update enterprise-web-search.md * Update enterprise-web-search.md
3.5 KiB
3.5 KiB
catalog_title, catalog_description, catalog_icon, catalog_tags
| catalog_title | catalog_description | catalog_icon | catalog_tags | ||
|---|---|---|---|---|---|
| Enterprise Web Search | Ground ADK agents with policy-compliant web search results | /integrations/assets/enterprise-web-search.png |
|
Enterprise Web Search tool for ADK
Supported in ADKPython v1.9.0TypeScript v1.5.0
The Google Cloud Enterprise Web Search grounds ADK agents with information from the web while maintaining enterprise compliance and source control. Designed for enterprise-grade workloads, this tool ensures grounding data aligns with organizational security and compliance policies.
!!! note "Enterprise Web Search vs. Agent Search"
Enterprise Web Search is distinct from
[Agent Search](https://adk.dev/integrations/agent-search/). While the Agent
Search queries indexed private data stores, Enterprise Web Search
retrieves compliant public web data.
!!! warning "Service Specific Terms"
When using the Enterprise Web Search tool, you are obligated to comply with Google's Service Specific Terms, which include properly displaying search suggestions and required Google logos in your UI.
Use cases
- Enterprise Grounding: Provide up-to-date web information to agents while maintaining organizational compliance standards.
- Controlled Web Access: Ensure agents query trusted web sources for research, market intelligence, or customer support tasks.
- Regulated Workflows: Deploy grounding capabilities in environments requiring strict auditability and data governance.
Prerequisites
- Access to Google Cloud Platform with Agent Search enabled.
- Configured GCP project with necessary permissions for Gemini models.
- The environment variable
GOOGLE_GENAI_USE_ENTERPRISE=TRUEmust be set. - The
google-adkpackage (Python) or@google/adkpackage (TypeScript) installed:
=== "Python"
```bash
pip install google-adk
```
=== "TypeScript"
```bash
npm install @google/adk
```
Use with agent
The following example demonstrates how to configure an ADK agent with the
pre-instantiated enterprise_web_search tool:
=== "Python"
```python
from google.adk.agents import Agent
from google.adk.tools import enterprise_web_search
root_agent = Agent(
model="gemini-flash-latest",
name="enterprise_search_agent",
instruction="Answer user questions accurately using enterprise-compliant web search results.",
tools=[enterprise_web_search],
)
```
=== "TypeScript"
```typescript
import { LlmAgent, ENTERPRISE_WEB_SEARCH } from "@google/adk";
const rootAgent = new LlmAgent({
model: "gemini-flash-latest",
name: "enterprise_search_agent",
instruction: "Answer user questions accurately using enterprise-compliant web search results.",
tools: [ENTERPRISE_WEB_SEARCH],
});
export { rootAgent };
```
Selection guidance
- Use standard Google Search for general-purpose applications requiring broad web coverage across any Gemini model.
- Use Enterprise Web Search when building enterprise agents that mandate compliance control, source auditing, and deployment on Gemini 2+ models.