AdsPower LocalAPI MCP Server
OfficialThe AdsPower LocalAPI MCP Server enables LLMs to interact with and manage AdsPower browser features through an API. Key capabilities include:
Browser Management: Open/close browsers using
userIdorserialNumber, create new browsers with custom configurations, update existing settings, and delete browsersGroup Operations: Create and update browser groups, move browsers between groups, and retrieve group lists
Configuration Control: Configure browser fingerprints, proxies, user agents, country settings, and other advanced parameters
Information Retrieval: Get lists of all browsers, currently opened browsers, groups, and available applications in the system
Allows creation and configuration of browser profiles with Android user agents
Supports configuration of Firefox browser profiles through the browser kernel configuration
Provides control over WebRTC settings in browser profiles, with options for disabling, forwarding, or using proxy/local connections
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@AdsPower LocalAPI MCP Server create a browser with random fingerprint for testing"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
AdsPower Skill & MCP & CLI
adspower-browser Skill
Claude Code/Codex/Cursor/OpenCode/Gemini GL
npx skills add https://github.com/adspower/adspower-browser --skill adspower-browserOpenClaw(小龙虾)
npx clawhub@latest install adspower-browser --forceHermes agent(爱马仕)
hermes skills install https://github.com/AdsPower/adspower-browser/blob/main/skills/adspower-browser --forceRelated MCP server: Puppeteer MCP Server
AdsPower CLI
The AdsPower CLI is completely isolated from the AdsPower client, runs independently, and does not share local caches/kernels, etc. However, the CLI also supports headless mode for operating the AdsPower client.
npm install -g adspower-browser # 安装
ads start -k <KEY> # 启动 adspower
ads stop # 停止 adspower
ads restart # 重启 adspower
ads status # 查询 adspower 运行状态
ads get-browser-list # 获取配置文件列表
... #查看 https://github.com/AdsPower/adspower-browser/blob/main/packages/adspower-browser/README.MDAdsPower client headless mode:
AdsPower Global:
Windows: "AdsPower Global.exe" --headless=true --api-key=your_api_key --api-port=50325
MacOS:"/Applications/AdsPower Global.app/Contents/MacOS/AdsPower Global" --args --headless=true --api-key=your_api_key --api-port=50325
Linux:adspower_global --headless=true --api-key=your_api_key --api-port=50325AdsPower MCP
Claude Code:
claude mcp add adspower-local-api -e PORT="50325" -e API_KEY="your_api_key" -- npx -y local-api-mcp-typescriptCodex:
codex mcp add --env PORT="50325" --env API_KEY="your_api_key" adspower-local-api -- npx -y local-api-mcp-typescriptOpenClaw:
openclaw mcp set adspower-local-api '{"command":"npx","args":["-y","local-api-mcp-typescript"],"env":{"PORT":"50325","API_KEY":"your_api_key"}}'Cursor: Go to Cursor Settings -> MCP -> Add New MCP Server, and then paste the following content:
{
"mcpServers": {
"adspower-local-api": {
"command": "npx",
"args": ["-y", "local-api-mcp-typescript"],
"env": {
"PORT": "50325",
"API_KEY": "your_api_key"
}
}
}
}OpenCode: update ~/.config/opencode/opencode.json
{
"$schema": "https://opencode.ai/config.json",
"mcp": {
"adspower-local-api": {
"command": "npx",
"args": ["-y", "local-api-mcp-typescript"],
"env": {
"PORT": "50325",
"API_KEY": "your_api_key"
}
}
}
}How to use?
Requirements
Node version 18 or greater
Version Compatibility
Some APIs are introduced incrementally by AdsPower client version.
If the client has not been upgraded to a version that includes a specific API, requests may return Not found.
When this happens:
Upgrade AdsPower client to the latest patch version first.
Then retry the same command/API.
Port Configuration
The port parameter specifies the AdsPower Local API port. Default is 50325.
Priority order:
Command-line argument
--port(highest priority)Environment variable
PORT(medium priority)Default value
50325(lowest priority)
Examples:
Using command-line argument:
{
"mcpServers": {
"adspower-local-api": {
"command": "npx",
"args": ["-y", "local-api-mcp-typescript", "--port", "50325"]
}
}
}Using environment variable:
{
"mcpServers": {
"adspower-local-api": {
"command": "npx",
"args": ["-y", "local-api-mcp-typescript"],
"env": {
"PORT": "50325"
}
}
}
}Development
# clone
git clone https://github.com/AdsPower/adspower-browser.git
cd adspower-browser
# install (monorepo)
pnpm install
# build all packages
pnpm run buildFor local MCP development, point Claude Desktop to the MCP package:
"mcpServers": {
"adspower-local-api": {
"command": "node",
"args": ["<Replace Your Project Path>/adspower-browser/packages/local-api-mcp/build/index.js"]
}
}Use AdsPower Cli in Docker
copy the /packages/adspower-browser/docker-compose.yml
cd packages/adspower-browser
docker-compose -f ./docker-compose.yml up -d
docker-compose exec adspower-cli /bin/bashAvailable Tools
27 toolsclick-elementC
Click the element
Inputs: selector (required).
close-browserC
Close the browser
Inputs: userId (required).
connect-browser-with-wsC
Connect the browser with the ws url
Inputs: wsUrl (required), userId, serialNumber.
create-browserD
Create a browser
Inputs: groupId (required), userProxyConfig (required), domainName, openUrls, +8 more.
create-groupC
Create a browser group
Inputs: groupName (required), remark.
delete-browserC
Delete the browser
Inputs: userIds (required).
drag-elementD
Drag the element
Inputs: selector (required), targetSelector (required).
evaluate-scriptD
Evaluate the script
Inputs: script (required).
fill-inputC
Fill the input
Inputs: selector (required), text (required).
get-application-listC
Get the list of applications
Inputs: size.
get-browser-listC
Get the list of browsers
Inputs: groupId, size, page, id, +3 more.
get-group-listC
Get the list of groups
Inputs: groupName, size, page.
get-opened-browserB
Get the list of opened browsers
Inputs: no named parameters.
get-page-htmlB
Get the html content of the page
Inputs: no named parameters.
get-page-visible-textB
Get the visible text content of the page
Inputs: no named parameters.
hover-elementD
Hover the element
Inputs: selector (required).
iframe-click-elementC
Click the element in the iframe
Inputs: selector (required), iframeSelector (required).
move-browserC
Move browsers to a group
Inputs: groupId (required), userIds (required).
open-browserD
Open the browser, both environment and profile mean browser
Inputs: serialNumber, userId, ipTab, launchArgs, +2 more.
open-new-pageC
Open a new page
Inputs: no named parameters.
press-keyD
Press the key
Inputs: key (required), selector.
screenshotC
Get the screenshot of the page
Inputs: savePath, isFullPage.
scroll-elementC
Scroll the element
Inputs: selector (required).
select-optionD
Select the option
Inputs: selector (required), value (required).
update-browserD
Update the browser
Inputs: userId (required), domainName, openUrls, cookie, +9 more.
update-groupC
Update the browser group
Inputs: groupId (required), groupName (required), remark.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 27 tools
Most tools have distinct purposes, such as create-browser vs. open-browser, but some overlap exists, like get-browser-list and get-opened-browser, which could cause confusion. The descriptions help clarify, but the boundaries between these list tools are not perfectly clear.
All tool names follow a consistent verb-noun pattern with hyphens, such as create-browser, get-browser-list, and navigate. There are no deviations in style or convention, making the naming highly predictable and readable throughout the set.
With 27 tools, the count is borderline high for a browser automation server, as it may feel heavy and complex for agents to navigate. However, given the domain's scope covering browser management, groups, and page interactions, it is not extreme but could benefit from consolidation.
The tool set provides comprehensive coverage for browser automation, including CRUD operations for browsers and groups, navigation, element interactions, and page content retrieval. There are no obvious gaps, and the surface supports full lifecycle management and interaction workflows.
Maintenance
Related MCP Connectors
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
Model Context Protocol server for the Apideck Unified API. Connect any MCP-compatible agent framework to 100+ accounting systems, HRIS platforms, file storage providers, and more through one integration. More information https://www.apideck.com/mcp-server
A Model Context Protocol server for Wix AI tools
Official Browserless MCP server: browser automation, scraping, screenshots and PDFs for AI agents.
Related MCP Servers
- AlicenseBqualityDmaintenanceA Model Context Protocol server that provides browser automation capabilities using Playwright. This server enables LLMs to interact with web pages, take screenshots, and execute JavaScript in a real browser environment.3215,256 npm5,681MIT
- AlicenseNot gradedqualityDmaintenanceA Model Context Protocol server that provides browser automation capabilities using Playwright, enabling LLMs to interact with web pages, take screenshots, and execute JavaScript in a real browser environment.19Apache 2.0
- AlicenseNot gradedqualityDmaintenanceA Model Context Protocol server that enables LLMs to automate web browsers using Puppeteer, allowing navigation, screenshots, form filling, and JavaScript execution in a real browser environment.2MIT
- AlicenseBqualityDmaintenanceA Model Context Protocol server that provides browser automation capabilities using Puppeteer, enabling LLMs to interact with web pages, take screenshots, and execute JavaScript in a real browser environment.728,579 npm1MIT