99tests MCP server
OfficialThis server has a remotely hosted endpoint. Open its connector listing for connection details, authentication requirements, and setup instructions:
99tests MCP server and plugin
Order a human-run, real-device test of your app or website from Claude, Cursor, VS Code, Cline, Gemini CLI or any MCP client, and get verified findings back as ready-to-paste fix prompts.
99tests has run crowdsourced testing since 2010: 20,000 registered testers across India on real phones and browsers, fixed prices shown before you pay, and a verified report within days. Docs and pricing: https://99tests.com/mcp-server
What it does
Your coding agent runs automated tests. 99tests adds what those cannot cover: people using your app on phones, laptops and browsers they own, on real networks. A reviewer verifies every finding before it reaches you.
The MCP server gives the agent five tools: list the packages, create a test job, check its status, fetch the verified report, and show a sample report.
The plugin for Claude and the extension for Gemini CLI add guidance on when and how to use those tools: show prices exactly as returned, ask before ordering, hand you the checkout link, and apply the report's fix prompts one at a time.
Related MCP server: Menso MCP Server
Endpoint
Remote server, Streamable HTTP, nothing to install:
https://99tests.com/api/mcpInstall
Claude Code plugin (the server plus the real-device-test skill):
/plugin install 99tests --marketplace 99tests/mcp-serverOn older Claude Code versions, run these two instead:
/plugin marketplace add 99tests/mcp-server
/plugin install 99tests@99testsClaude Code, server only:
claude mcp add --transport http 99tests https://99tests.com/api/mcpGemini CLI extension:
gemini extensions install https://github.com/99tests/mcp-serverCursor (mcp.json):
{
"mcpServers": {
"99tests": {
"url": "https://99tests.com/api/mcp"
}
}
}VS Code (.vscode/mcp.json):
{
"servers": {
"99tests": {
"type": "http",
"url": "https://99tests.com/api/mcp"
}
}
}Cline (cline_mcp_settings.json):
{
"mcpServers": {
"99tests": {
"type": "streamableHttp",
"url": "https://99tests.com/api/mcp"
}
}
}Other MCP clients: add a remote MCP server with the URL above. Notes for AI agents that install it: llms-install.md.
Use it
Ask your agent, for example:
"List the 99tests packages and tell me which one fits a release check of my Android app on budget phones."
"Create a Standard 99tests job and give me the checkout link."
"Check my 99tests job. If the report is ready, apply the fix prompts one by one."
"Show me a sample 99tests report."
Tools
Tool | What it does |
| Lists test packages with scope, device counts and prices |
| Creates a test job and returns a browser checkout link |
| Returns the job's stage and progress |
| Returns the verified findings for a delivered job, as fix prompts |
| Returns a sample report so an agent can see the output format |
How a cycle runs
The agent calls
create_test_jobwith a package and your email address, and gives you a checkout link.You pay in your browser. No account, no sales call. Then you submit the build link or URL, your test cases and the target devices.
Testers run the sessions on real devices; a reviewer verifies every finding.
get_reportreturns the findings. Each one includes steps, device, OS, screenshots and a fix prompt the agent can act on.
Data
The server receives only what a tool call carries: the package you choose, the email address you give for the invoice, status link and report, an optional country code, and your job's status link or token. The plugin and the extension send nothing else and store nothing. Your source code and files are not sent. Builds and test cases go to 99tests in the submit step after checkout, not through the agent.
Privacy policy: https://99tests.com/privacy
Terms: https://99tests.com/terms
Links
Server docs: https://99tests.com/mcp-server
Pricing: https://99tests.com/pricing
Sample report: https://99tests.com/sample-report/
Free QA tools (website checker, Play Store bug scanner, test case library): https://99tests.com/tools
India App Quality Report: https://99tests.com/research/india-app-quality-report-q4-2026
Support
hello@99tests.com · Issues on this repository are welcome for client-compatibility bugs.
Licence
This repository (docs, manifests and the skill) is MIT licensed. The hosted service is provided under the 99tests client terms: https://99tests.com/terms
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Human QA on real phones and browsers for your coding agent. Pass/fail per step, with screenshots.
Usability tests with real people, ordered from your coding agent. Results within 24 hours.
Agents that test your web and mobile app like real users, on every pull request.
Agents that test your web and mobile app like real users, on every pull request.
Related MCP Servers
- AlicenseAqualityDmaintenanceEnables AI agents to perform comprehensive web application testing including visual, functional, performance, accessibility, and SEO analysis using browser automation without requiring API keys.7MIT

Menso MCP Serverofficial
AlicenseNot gradedqualityCmaintenanceEnables AI clients to launch real purchase or signup test runs against a live public URL and then fetch run status, scored findings with friction points and suggested fixes, and a public replay of every step taken. It works over Streamable HTTP with an API key, so no installation is needed in Claude Code, Cursor, Windsurf or any compatible client.2 npmMIT- AlicenseNot gradedqualityCmaintenanceEnables MCP clients to launch AI persona tests against public web app URLs in a real browser, then retrieve status, findings, screenshots, and combined fix prompts for coding agents.MIT

TestOptim MCP serverofficial
AlicenseNot gradedqualityCmaintenanceConnects MCP clients such as Claude Code, Cursor and VS Code to a hosted QA service so the assistant can start and monitor test runs on a web app, explore its pages and flows, and read the bugs those runs uncover. Access is scoped by API key permissions, letting read-only keys list projects and issues while write-enabled keys trigger or stop runs.MIT