👁 Image
STAF.UI.API 4.4.2

STAF - Simple Test Automation Framework

C# Selenium | .NET Selenium Framework | MCP Server for Selenium C# | Selenium WebDriver C# | Test Automation .NET | UI Testing C# | API Testing .NET | MSTest Selenium

👁 NuGet
👁 .NET 10

STAF is a production-ready .NET test automation framework for Selenium-based UI testing, API testing, and Excel validation. It provides base classes, HTML reporting, parallel execution, and optional database and accessibility support. The framework is distributed as the STAF.UI.API NuGet package and targets .NET 10 with MSTest.

Note: This release targets .NET 10. Projects consuming STAF.UI.API must use .NET 10 or above.

Table of Contents

• AI-assisted development and MCP (implementation repo)
• Overview
• Features
• Key Components
• Getting Started
• Configuration
• Reporting
• Parallel execution and thread safety
• Upgrading and migration
• Usage Guide
• License

AI-assisted development and MCP (implementation repo)

This repository contains the STAF.UI.API framework source and ships the NuGet package. It does not include the MCP server executable or editor-specific AI configuration—those live in the official implementation / sample repository so you can pair the package with real projects and tooling.

Use this together with the implementation project:

STAF.Selenium.Tests is the official template that references STAF.UI.API and adds:

• MCP (Model Context Protocol) server for Selenium + STAF — a self-contained agent you can attach to Cursor, VS Code, or Visual Studio (Copilot agent mode) to drive browsers and generate tests from natural language.
• Editor rules — e.g. .cursor/rules, .github/copilot-instructions.md, and repo-level .mcp.json so assistants follow STAF patterns (TestBaseClass, FindAppElement, ReportResult, etc.).
• Working samples for UI, API, Excel, database, reporting, accessibility, and parallel runs.

Clone the implementation repo to run MCP, copy its config into your own solution, or start new suites from the template while keeping STAF.UI.API as the shared framework dependency:

git clone https://github.com/sooraj171/STAF.Selenium.Tests
cd STAF.Selenium.Tests
dotnet restore
dotnet build

For MCP setup (paths to MCPAgent/publish/..., tool list, troubleshooting), see MCPAgent/README.md and the main README in STAF.Selenium.Tests.

Overview

STAF streamlines automated testing for web applications and APIs using Selenium WebDriver and C#. It supports:

Features

Key Components

Test and Page Base Classes

Browser and Driver

Reporting

WebDriver Extensions

Excel, Database & Accessibility

Getting Started

Prerequisites

• .NET 10 SDK
• Visual Studio 2022 (or later) or VS Code with C# extension
• Chrome or Edge (for UI tests)
• MSTest (included via package reference)

Install

1. Add the NuGet package to your test project:

   dotnet add package STAF.UI.API
   

   Or use the STAF.Selenium.Tests implementation project for samples, runsettings, and optional MCP / AI tooling (see AI-assisted development and MCP).

2. Configure run settings: Test > Configure Run Settings > Select Solution Wide runsettings File > choose testrunsetting.runsettings.

3. Build and run tests from Test Explorer or CLI: dotnet test.

Sample UI Test

[TestMethod]
public void TestLogin()
{
 var element = FindAppElement(By.Id("loginButton"), "Login button");
 ReportResult.ReportResultPass(driver, TestContext, "Login", "Login button is visible");
}

Sample API Test

[TestMethod]
public void TestAPIStatus()
{
 ReportResultAPI.ReportResultPass(TestContext, "API", "Status check passed");
}

Configuration

Run settings (testrunsetting.runsettings)

• TestRunParameters: browser (e.g. chrome), driverPath, url, optional email/smtp settings
• headless: set to true for CI/agents (Chrome/Edge headless is applied after your SetChromeOptions / SetEdgeOptions overrides)
• killOrphanChromedrivers: default false. Set to true only on isolated agents if you still need assembly-level cleanup that terminates all chromedriver processes (avoid on shared build machines)
• MSTest: Parallelize (e.g. Workers=4, Scope=MethodLevel)
• ResultsDirectory: e.g. .\TestResults

appsettings.json

AppConfig.GetConfig() looks for appsettings.json in this order: AppContext.BaseDirectory, the executing assembly directory, then Directory.GetCurrentDirectory(). The first match wins.

• ConnectionStrings: e.g. DefaultConnection for DbHelper
• Email: SmtpHost, SmtpPort, UseDefaultCred, Username, Password (optional)

Overriding browser options

protected override ChromeOptions SetChromeOptions()
{
 var options = new ChromeOptions();
 options.AddArguments("start-maximized");
 return options;
}

Use runsetting headless=true for pipelines; the framework adds headless arguments after this method runs, so existing overrides stay valid.

Reporting

In-test step reporting

ReportResult.ReportResultPass(driver, TestContext, "ModuleName", "Description");
ReportResult.ReportResultFail(driver, TestContext, "ModuleName", "Description", "exception text");

// API tests (no driver)
ReportResultAPI.ReportResultPass(TestContext, "ModuleName", "Description");

Element assertions with reporting

element.ReportElementExists(driver, TestContext, testName, "Element exists", ProdceedFlag: true);
element.ReportElementIsDisplayed(driver, TestContext, testName, "Element is displayed", ProdceedFlag: true);

Assembly summary

After all tests, AssemblyCleanup merges per-test HTML fragments using a file-based accumulator, appends them to ResultTemplate.html, and copies the outcome to ResultTemplateFinal.html next to the test run output. Those files are created under TestContext.TestRunDirectory when the test host provides it (typical under TestResults); otherwise the framework falls back to the test assembly base directory.

Fail screenshots are written next to the per-test HTML report with unique names; the report still embeds them as base64 for the same in-browser experience as before.

Programmatic HTML (TestReportGenerator)

TestReportGenerator.GenerateTestReport(@"C:\out\my_report.html", myTestResultData);

Use a unique filePath when multiple processes or tools might generate reports at the same time.

Parallel execution and thread safety

• Per-test state (result HTML path, failure flag) uses AsyncLocal so parallel MSTest workers do not overwrite each other. The framework still mirrors failFlag and per-test path keys to environment variables for backward compatibility in sequential runs.
• HTML append uses a lock per report file path, not one global lock, so different tests can write their own reports concurrently.
• ReportResult / ReportResultAPI resolve the report file via the active test context first, then fall back to Environment.GetEnvironmentVariable(TestContext.TestName) if you use a custom startup path.

Upgrading and migration

When updating from older STAF.UI.API builds, check the following:

Usage Guide

Excel comparison

var excel = new ExcelDriver();
ExcelCompareStatus res = excel.CompareFiles("path/Excel1.xlsx", "path/Excel2.xlsx", sheetIndex1: 1, sheetIndex2: 1);
if (res.IsMatching) return true;

Accessibility (Axe)

var axe = new AxeAccessibility(driver);
axe.AnalyzePage(); // Full page scan
axe.AnalyzePageAndSaveHtml("report.html"); // Run and save HTML report
axe.AnalyzeCssSelector("#main"); // Scoped scan
axe.AnalyzeWithConfigurator(b => b.Include(".modal")); // Custom configuration

Database (DbHelper)

Configure connection in appsettings.json under ConnectionStrings:DefaultConnection, then use:

DbHelper.VerifyConnection();
var value = DbHelper.ExecuteScalar<int>("SELECT COUNT(*) FROM Users");

License

This project is licensed under the MIT License.

Author: Sooraj Ramachandran
Copyright (c) 2026 Sooraj Ramachandran. All rights reserved.

This software is provided "as is", without warranty of any kind. See the for full terms.