George Weale
d9e930e823
* docs(integrations): fix unresolvable imports and stale API claims * docs: apply style pass and drop out-of-scope import cleanup * docs(integrations): address review feedback on gcs, cloud-trace, reflect-and-retry Restore the gcs_ tool name prefixes in the GCS tool tables, since both toolsets set tool_name_prefix="gcs" and the tables list names as the model sees them. Use the current Agent Platform SDK name in cloud-trace prose, make the reflect-and-retry failure description language-neutral for Python and Go, and drop the redundant re-export clause. * docs(gcs): note that tool_filter matches unprefixed tool names Tool filtering runs inside get_tools() against the unprefixed name, and get_tools_with_prefix() applies the gcs_ prefix afterwards, so the names in the tables are not the names tool_filter expects. * docs(computer-use): drop unused Gemini and override imports --------- Co-authored-by: Kristopher Overholt <koverholt@google.com>
3.4 KiB
catalog_title, catalog_description, catalog_icon, catalog_tags
| catalog_title | catalog_description | catalog_icon | catalog_tags | |
|---|---|---|---|---|
| Computer Use | Operate computer user interfaces using Gemini models | /integrations/assets/gemini-spark.svg |
Gemini API Computer Use tool for ADK
The Computer Use Toolset allows an agent to operate a user interface of a computer, such as browsers, to complete tasks. This tool uses a specific Gemini model and the Playwright testing tool to control a Chromium browser and can interact with web pages by taking screenshots, clicking, typing, and navigating.
For more information about the computer use model, see Gemini API Computer use or the Agent Platform API Computer use.
!!! example "Preview release" The Computer Use model and tool is a Preview release. For more information, see the launch stage descriptions.
Setup
You must install Playwright and its dependencies, including Chromium, to be able to use the Computer Use Toolset.
??? tip "Recommended: create and activate a Python virtual environment"
Create a Python virtual environment:
```shell
python3 -m venv .venv
```
Activate the Python virtual environment:
=== "Windows CMD"
```console
.venv\Scripts\activate.bat
```
=== "Windows Powershell"
```console
.venv\Scripts\Activate.ps1
```
=== "MacOS / Linux"
```bash
source .venv/bin/activate
```
To set up the required software libraries for the Computer Use Toolset:
- Install Python dependencies:
pip install termcolor==3.1.0 pip install playwright==1.52.0 pip install browserbase==1.3.0 pip install rich - Install the Playwright dependencies, including the Chromium browser:
playwright install-deps chromium playwright install chromium
Use the tool
Use the Computer Use Toolset by adding it as a tool to your agent. When you
configure the tool, you must provide a implementation of the BaseComputer
class which defines an interface for an agent to use a computer. In the
following example, the PlaywrightComputer class is defined for this purpose.
You can find the code for this implementation in playwright.py file of the
computer_use
agent sample project.
from google.adk import Agent
from google.adk.tools.computer_use.computer_use_toolset import ComputerUseToolset
from .playwright import PlaywrightComputer
root_agent = Agent(
model='gemini-2.5-computer-use-preview-10-2025',
name='hello_world_agent',
description=(
'computer use agent that can operate a browser on a computer to finish'
' user tasks'
),
instruction='you are a computer use agent',
tools=[
ComputerUseToolset(computer=PlaywrightComputer(screen_size=(1280, 936)))
],
)
For a complete code example, see the computer_use agent sample project.