google-adk-builtin-tools

name	google-adk-builtin-tools
description	Catalog of all ADK built-in tools. Use when looking for pre-built tools to add to agents — search, memory, artifacts, transfer, grounding, and user interaction tools.

Google ADK — Built-in Tools Catalog

Import

from google.adk.tools import (
    google_search,
    exit_loop,
    load_memory,
    preload_memory,
    load_artifacts,
    get_user_choice,
    transfer_to_agent,
    url_context,
    google_maps_grounding,
    enterprise_web_search,
    LongRunningFunctionTool,
    VertexAiSearchTool,
    DiscoveryEngineSearchTool,
    ExampleTool,
)

Search & Grounding Tools

Tool	Description	Usage
google_search	Web search via Google Search	tools=[google_search]
url_context	Fetches and reads web page content	tools=[url_context]
google_maps_grounding	Grounds responses with Google Maps data	tools=[google_maps_grounding]
enterprise_web_search	Enterprise web search	tools=[enterprise_web_search]
VertexAiSearchTool	Vertex AI Search datastore	See below
DiscoveryEngineSearchTool	Discovery Engine search	See below

google_search

from google.adk.agents import Agent
from google.adk.tools import google_search

agent = Agent(
    name="researcher",
    model="gemini-2.5-flash",
    instruction="Research topics using web search.",
    tools=[google_search],
)

VertexAiSearchTool

from google.adk.tools import VertexAiSearchTool

vertex_search = VertexAiSearchTool(
    data_store_id="projects/PROJECT/locations/LOCATION/collections/default_collection/dataStores/DATASTORE_ID",
)

agent = Agent(
    name="knowledge_agent",
    model="gemini-2.5-flash",
    instruction="Answer questions using the knowledge base.",
    tools=[vertex_search],
)

Memory Tools

Tool	Description	Usage
load_memory	On-demand memory recall (agent calls it)	tools=[load_memory]
preload_memory	Auto-loads relevant memories before each turn	tools=[preload_memory]

from google.adk.agents import Agent
from google.adk.tools import load_memory, preload_memory

agent = Agent(
    name="remembering_agent",
    model="gemini-2.5-flash",
    instruction="Remember past conversations. Use load_memory when relevant.",
    tools=[load_memory],
)

Artifact Tools

Tool	Description	Usage
load_artifacts	Loads saved artifacts for the agent to view	tools=[load_artifacts]

from google.adk.agents import Agent
from google.adk.tools import load_artifacts

agent = Agent(
    name="viewer",
    model="gemini-2.5-flash",
    instruction="Help users view their saved files.",
    tools=[load_artifacts],
)

Agent Transfer Tools

Tool	Description	Usage
transfer_to_agent	Explicitly transfers control to a named agent	tools=[transfer_to_agent]

from google.adk.agents import Agent
from google.adk.tools import transfer_to_agent

agent = Agent(
    name="router",
    model="gemini-2.5-flash",
    instruction="Transfer to specialist agents as needed.",
    tools=[transfer_to_agent],
    sub_agents=[specialist_a, specialist_b],
)

Loop Control Tools

Tool	Description	Usage
exit_loop	Breaks out of a LoopAgent iteration	tools=[exit_loop]

from google.adk.agents import Agent, LoopAgent
from google.adk.tools import exit_loop

refiner = Agent(
    name="refiner",
    model="gemini-2.5-flash",
    instruction="Refine until quality is good, then call exit_loop.",
    tools=[exit_loop],
)

loop = LoopAgent(name="loop", sub_agents=[refiner], max_iterations=5)

User Interaction Tools

Tool	Description	Usage
get_user_choice	Prompts user to choose from options	tools=[get_user_choice]

from google.adk.agents import Agent
from google.adk.tools import get_user_choice

agent = Agent(
    name="interactive",
    model="gemini-2.5-flash",
    instruction="When uncertain, ask the user to choose.",
    tools=[get_user_choice],
)

Long-Running Tools

Wraps a function tool that takes a long time (avoids timeout):

from google.adk.tools import LongRunningFunctionTool

def slow_analysis(data: str) -> str:
    """Performs deep analysis that may take minutes.

    Args:
        data: The data to analyze.
    """
    result = run_expensive_computation(data)
    return result

agent = Agent(
    name="analyzer",
    model="gemini-2.5-flash",
    instruction="Analyze data thoroughly.",
    tools=[LongRunningFunctionTool(func=slow_analysis)],
)

Example Tool (Few-Shot)

Provides examples of input/output to guide the model:

from google.adk.tools import ExampleTool

example_tool = ExampleTool(
    name="format_example",
    examples=[
        {"input": "john doe", "output": "John Doe"},
        {"input": "jane smith", "output": "Jane Smith"},
    ],
)

Key Rules

• Built-in tools are imported from google.adk.tools
• Most built-in tools are singleton instances — pass them directly (no instantiation)
• VertexAiSearchTool and DiscoveryEngineSearchTool need instantiation with config
• LongRunningFunctionTool wraps your function — don't pass it directly to tools
• Memory tools require a memory_service configured on the Runner
• exit_loop only works inside a LoopAgent's sub-agents
• transfer_to_agent is implicit in multi-agent systems but can be explicit

Related Skills

• google-adk-function-tool — Creating custom function tools
• google-adk-memory — Memory service configuration (for load_memory/preload_memory)
• google-adk-artifacts — Artifact service (for load_artifacts)
• google-adk-workflow-agents — LoopAgent (for exit_loop)