The SQAI Suite MCP (Model Context Protocol) server acts as a centralized intelligence layer for your development environment. It connects your code, user stories, and test management tools into a single, real-time understanding of the system under development.
Added Value: A Context-Aware Bridge
By connecting the SQAI MCP to your IDE, you create a direct link between your coding environment and your broader project data:
Unified Intelligence: It connects code repositories, user stories, and documentation so your AI understands the relationship between different artifacts.
Automated Traceability: Easily perform tasks like automated test generation, impact analysis, or documentation creation without leaving your editor.
Continuous Updates: As you change your code or update a user story, the MCP enables you to keep the system model updated, ensuring your AI assistant is never working with stale information.
1. Generate Your API Key
You need an authentication key to securely connect your IDE to the SQAI Platform.
Log in to the SQAI Platform.
Go to the MCP Configuration page: app.sqai-suite.com/mcp.
Generate and copy your unique API Key.
2. Setup Instructions
You can configure this integration globally for all your work or locally for a specific project team (see 2.1)
Security Note: If you choose the local project method (see section 2.1), it is safe to check the mcp.json file into Git. Your API key is not stored in this file; it is stored securely in your private local settings and prompted only upon server start.
2.1 Choose Your Method
It is advised to only add the config in one place (locally per project OR globally), otherwise the MCP server may appear twice. If you install it per project, you need one mcp.json file per project.
The advantage of project-level installation is that you can check the configuration into your repository so all team members share the same setup. The API key is never included in this JSON; the inputs array refers to a locally encrypted and stored input unique to each user, making it safe to commit to version control.
Local (Per Project): Create a file at
.vscode/mcp.jsonin your project root.Global (All Projects): Press
CTRL+SHIFT+P, type "MCP: Open Global Config", and select it.
2.2 Configuration Snippet
Paste the following into your mcp.json file:
{
"inputs": [
{
"type": "promptString",
"id": "sqai-api-key",
"description": "API key for SQAI MCP",
"password": true
}
],
"servers": {
"SQAI-Suite": {
"type": "http",
"url": "https://api.sqai-suite.com/mcp",
"headers": {
"Authorization": "Bearer ${input:sqai-api-key}"
}
}
}
}
3. How to Use
Open Copilot Chat: Use the chat icon or
CTRL + SHIFT + I.Enable Agent Mode: Ensure the chat is in Agent mode to allow it to communicate with the MCP server.
Authenticate: When prompted on the first run, paste the API key you generated in Step 1.
Interact: Ask Copilot to utilize your SQAI data. Example:
"Based on the user story for this feature, generate the necessary test cases and sync them to our test management tool."
You'll notice that it will automatically recognize SQAI's MCP Capabilties.
In case you would like an overview of which capabilties apply to your SQAI MCP, you can click the tools icon and get the full list with descriptions:
4. Troubleshooting
Issue | Potential Cause | Solution |
Server not appearing | Incorrect Chat mode. | Ensure Copilot Chat is set to Agent mode. Standard chat modes do not support MCP. |
Repeated auth prompts | Config conflict. | Check for |
Connection failed | Network/Proxy restrictions. | Verify access to |
Server startup error | Incorrect JSON or key. | Run "MCP: List Servers" from the Command Palette, select SQAI-Suite, and click Show Output for detailed logs. |
Unresponsive tools | Cache issues. | Run the "MCP: Reset Cached Tools" command from the Command Palette to refresh the server state. |