MCP Server · open source on npm

Autonomous testing
in your AI coding agent

Your coding agent runs the tests and files the bugs, straight into GitHub. Connect Wopee.io to generate tests, dispatch testing agents, and check results without leaving your editor.

Let your coding agent set it up

Claude Code
Cursor
Copilot
Codex

Set me up for Wopee.io: follow https://wopee.io/agent-setup/prompt.md, show me the plan first, and ask before changing anything.

Paste it into your agent’s chat. It shows you the plan before it changes anything.What the prompt does

No coding agent? Start for free in the browser → · No credit card required

npm package•15 MCP tools•Node.js 20.18+

AI Agent + Wopee.io MCP

mcp connected

Read the docs →New to MCP? Free tutorial →

Short answer

Is there an MCP server for testing web apps?

Yes. Wopee.io's MCP server, wopee-mcp on npm, connects Claude Code, Cursor, GitHub Copilot, OpenAI Codex, Devin Desktop or any other MCP client to Wopee.io's AI testing agents. From the chat, your coding agent can generate tests, start test runs against your web app, fetch the results and file GitHub issues. It is for testing your web app through MCP, not for testing an MCP server.

Last checked: 9 November 2026

Your AI can code. But who tests the output?

Your CI catches technical errors but misses broken user flows
Writing test scripts manually slows down every release
You leave your editor to open a separate testing dashboard
Testing dozens of user flows manually doesn't scale

Wopee.io MCP bridges the gap. Test from where you code.

Up and running in 3 minutes

Paste one line, add your API key, and start testing

chat

Set me up for Wopee.io: follow https://wopee.io/agent-setup/prompt.md, show me the plan first, and ask before changing anything.

Paste

One line in your agent's chat

Claude Code, Cursor, Copilot or Codex reads the public setup file, checks Node.js 20.18+, and shows you the plan. Nothing changes until you say yes.

terminal

# Commander: More › Settings › API Keys
export WOPEE_API_KEY="…"
export WOPEE_PROJECT_UUID="…"

Connect

Your key stays with you

Your agent adds one wopee server that reads the key from your environment or a git-ignored .env file. You type the key yourself; it never goes into the chat or a committed file.

conversation

"Generate stories for login and test them"

  wopee_generate_artifact           3 user stories
  wopee_dispatch_agent              running 8 cases
  wopee_fetch_executed_test_cases   7 passed, 1 failed

Test

Natural language, real results

Ask your AI to test any flow. It generates stories, dispatches agents, and explains findings, all in natural language.

Plug into the AI tools you already use

The Model Context Protocol (MCP) is the open standard that lets AI agents use external tools. Any MCP-compatible client works with Wopee.io.

Claude CodeCLI agent by Anthropic
CursorAI-first code editor
VS CodeWith GitHub Copilot
OpenAI CodexCLI, IDE and ChatGPT app
Devin DesktopFormerly Windsurf
Claude DesktopAI agent by Anthropic
Any MCP clientOpen standard

Setup guide for each client →

Complete testing workflow

From crawling your app to filing GitHub issues, 15 tools, all through conversation

Dispatch Analysis Crawl

wopee_dispatch_analysis

Create a suite and dispatch a crawling agent to explore your web app, discover pages, and map the structure.

Generate Test Artifacts

wopee_generate_artifact

Generate user stories, test cases, test steps, and Playwright code from your app context.

Dispatch Testing Agent

wopee_dispatch_agent

Execute test cases with an autonomous AI agent that opens a real browser, follows test steps, and captures screenshots.

Fetch Execution Results

wopee_fetch_executed_test_cases

Retrieve pass/fail results, agent reports, and code reports for executed test cases.

Check Recent Runs

wopee_fetch_recent_executions

Get the latest test executions for your project (status, reports, and pass/fail) without hunting for suite UUIDs.

File GitHub Issues

wopee_create_github_issue

Turn test failures into GitHub issues in your connected repository, title, Markdown body, and labels.

Fetch Test Artifacts

wopee_fetch_artifact

Retrieve generated artifacts, user stories, test cases, Playwright code, and app context.

Update Test Artifacts

wopee_update_artifact

Replace and refine user stories, test cases, or Playwright code after reviewing them.

Create Suite

wopee_create_blank_suite

Create a new empty analysis suite as the starting point for your testing workflow.

List Suites

wopee_fetch_analysis_suites

List all analysis suites in your project with their UUIDs, types, and statuses.

Fetch Run-Time Variables

wopee_fetch_variables

Read the run-time variables that drive analysis and agent runs, at project level or per analysis suite.

Update Run-Time Variables

wopee_update_variables

Add or overwrite run-time variables at project or suite level, existing keys you don't list are preserved.

Fetch Test Inventory

wopee_fetch_test_inventory

The complete picture: every authored test case joined with its latest execution status, including never-run ones.

Post to Project Chat

wopee_send_chat_message

Post status updates and notes to your project's chat room so the whole team sees what the agent is doing.

Read Chat History

wopee_read_chat_history

Pull recent messages from your project's chat room for context before the agent acts.

Read the docs →

Video thumbnail: AI QA agent in Cursor using the Wopee.io, Playwright and Atlassian MCP servers

See it in action

An AI QA agent in Cursor combines the Wopee.io, Playwright and Atlassian MCP servers to retest a Jira bug and post the report back.

Watch the 4:36 video →

Available on GitHub and npm

The Wopee.io MCP server is published on npm and the source code is available on GitHub. Install in seconds, inspect the code, and contribute if you'd like.

Star on GitHub View on npm

Free tutorial: MCP for testers →

wopee-mcp

15 tools

# analyze & generate

wopee_dispatch_analysisCrawl and analyze your app

wopee_generate_artifactGenerate test artifacts

wopee_create_blank_suiteCreate a new suite

# run tests

wopee_dispatch_agentExecute test cases

wopee_fetch_executed_test_casesGet execution results

wopee_fetch_recent_executionsCheck recent runs

# suites, artifacts & variables

wopee_fetch_analysis_suitesList all suites

wopee_fetch_test_inventoryComplete test inventory

wopee_fetch_artifactRetrieve test artifacts

wopee_update_artifactUpdate test artifacts

wopee_fetch_variablesRead run-time variables

wopee_update_variablesUpdate run-time variables

# collaborate & report

wopee_create_github_issueFile bugs on GitHub

wopee_send_chat_messagePost to project chat

wopee_read_chat_historyRead chat history

Join the visionaries shaping the future

Trusted by the technology leaders & innovators

“Wopee showcased its potential to save us valuable time and effort in maintaining the visual integrity of our application”

Livesport, The fastest sports information and scores provider from 35+ sports to 100M people worldwide (60+ web and mobile apps)

Martin Šimon

Test Automation Lead

Bring autonomous testing to your AI workflow

Set up in minutes, start testing immediately.

No credit card required

Autonomous testing in your AI coding agent