Evaficy MCP for Evaficy Smart Test QA Platform
Bring your QA projects, test cases and run results into your AI assistant. Ask questions about your testing and generate new test cases straight from your editor.
What Is the Evaficy MCP?
The Model Context Protocol (MCP) is an open standard that lets AI assistants connect to external tools and data. An MCP server describes what it can do, and the assistant calls it whenever a question needs that information.
Evaficy MCP is the MCP server for Evaficy Smart Test. Once it is connected, your assistant can look up your projects, read test case scenarios, check test run results and generate new test cases, all in the same conversation where you write code. It is free, open source under the MIT license, and published on npm as evaficy-mcp.
What You Can Do With It
The MCP gives developers and testers quick access to QA context at the moment they need it.
Understand a project at a glance
Ask which projects you belong to, who the members are, and which environments are configured, all without switching to the browser.
Keep test cases next to the code
While building or fixing a feature, pull up the matching scenario with its steps, test data and validation status so the implementation lines up with what QA will check.
Check run results before a release
Find the latest test run for a project and get a summary of passed, failed and blocked test cases, plus the execution timeline.
Generate test cases from a requirement
Describe a requirement in plain language and get around 7 AI-generated test cases back. Review them as a preview, or ask for them to be saved as a new scenario in your project.
Available tools
You never need to call these by name; your assistant chooses them for you. They are listed here so you know exactly what the server can access.
Projects you own or belong to.
One project with members, environments and client.
Test case scenarios in a project.
A scenario with steps, test data and validation status.
Test runs, filterable by project and status.
Per-test-case results and the condensed run report.
AI test cases from a requirement, optionally saved as a scenario.
How It Works
You ask in plain language
For example, "What failed in the last Checkout run?" in Claude Code, Cline, Cursor or another MCP-capable client.
Your AI client picks the tool
It chooses the matching Evaficy tool and fills in the parameters. Project names are resolved to IDs automatically by listing your projects first.
Evaficy MCP calls the platform
The server runs locally on your machine and talks to app.evaficy.com using your Personal Access Token.
The answer comes back in your chat
Your assistant summarizes the data for you. Anything you choose to save, like a generated scenario, shows up in the Evaficy app straight away.
Get Started in Three Steps
Create a Personal Access Token
Log in to Evaficy Smart Test and open Settings → API Tokens → Create Token. Give it a name you will recognize, such as "Evaficy MCP - laptop". The token is shown once, so copy it right away. API tokens are available on every plan.
Add the server to your AI client
Register evaficy-mcp in your client and pass the token as the EVAFICY_TOKEN environment variable. The package runs through npx, so there is nothing to install globally. Node.js 18 or newer is required.
Ask your first question
Start with "List my Evaficy projects." It confirms the connection works and gives your assistant the project names and IDs it needs for follow-up questions.
Claude Code
claude mcp add evaficy -e EVAFICY_TOKEN=<your-token> -- npx evaficy-mcp
Cline, Cursor and other clients with a JSON config
{
"mcpServers": {
"evaficy": {
"command": "npx",
"args": ["evaficy-mcp"],
"env": {
"EVAFICY_TOKEN": "<your-token>"
}
}
}
}Treat your token like a password
A token carries the same permissions as the account that created it. Keep it out of source control, and revoke it from Settings → API Tokens if it is ever exposed. For extra safety, tokens are never accepted for login, billing or administration, and a token cannot be used to create further tokens.
Example Prompts
Talk to your assistant the way you would talk to a teammate. Here are a few prompts to try once the server is connected.
List my Evaficy projects.
Show me the test cases in the "Checkout Flow" scenario of the Web Shop project.
Summarize the most recent completed test run in Web Shop. Which test cases failed or were blocked?
Generate functional test cases for the Checkout page in Web Shop. Requirement: "Users can apply a discount code before payment and see the updated total." Save them as a new scenario called "Discount Code at Checkout".
Good to Know
Six of the seven tools are read-only
Listing and reading projects, test cases and runs never changes anything in your account. Only generate_test_cases creates something, and it saves a new scenario only when you ask it to. Otherwise it returns the generated test cases as a preview for you to review.
Generating test cases uses your AI quota
Each generation counts toward the same quota as generating in the app: 5 generations in total on the Basic plan, 200 per month on Advanced, and 500 per month on Enterprise. Use the preview to refine your requirement before saving.
Longer generations can be resumed
If a generation is still running after about 90 seconds, the server hands back a job ID. Ask your assistant to resume it and the results are picked up where they left off, as long as you do so within 10 minutes.
For tips on writing requirements that produce strong test cases, see Writing Effective Inputs for AI Test Case Generation. Full configuration details live in the project README on GitHub.
Related guides
Bring your QA data into your AI assistant
Create an Evaficy Smart Test account, add a Personal Access Token, and connect the MCP in a couple of minutes.
Start your trial