• LOGS
  • ANALYSIS
  • REPORTS
  • HUMAN

How to debug failed test sessions with your AI

Connect BrowserStack to your agent to pull logs and session data without leaving your code editor. Once connected, your AI handles the heavy lifting of log retrieval and error identification.

Ask AI about this page

Short answer

How can I debug BrowserStack failures in my IDE?

Your AI uses the BrowserStack MCP to fetch specific session data and raw execution logs. It then analyzes those logs to pinpoint the exact cause of a failure. You get a clear explanation of the error right in your chat window.

Debugging outcomes

Where your debugging data lands.

The AI organizes the raw data from your failed runs into actionable insights.

LOGS

Raw log retrieval

The AI pulls the exact text from the failed session. You get the raw output needed to see what actually happened on the browser.

ANALYSIS

Error identification

The agent scans the logs for stack traces or timeout messages. It identifies the specific line or condition that caused the crash.

REPORTS

Summary of failures

The AI compiles the findings into a concise summary. This helps you understand if the issue is a flaky test or a real bug.

HUMAN

Manual fix required

The AI points out the broken code or configuration. You decide how to patch the test or the application logic.

The workflow

What your AI does when a test fails.

Instead of hunting through the BrowserStack dashboard, you let your agent pull the data for you.

  1. Locate the session

    The agent identifies the specific failed run you are asking about.

    get_session

  2. Fetch execution logs

    The AI pulls the raw text logs from that specific session to see the error details.

    get_session_logs

  3. Parse error data

    The agent reads through the log files to find stack traces and error messages.

    get_session_logs

  4. Report the cause

    The AI explains why the test failed based on the retrieved log data.

    get_session_logs

Try it

Copy these to start.

Use these prompts to kick off the debugging process.

Starting points

Replace the session IDs and names with your actual BrowserStack data.

BrowserStack Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_QfMSmaBJhbb00bXFG15oIucKmIgIGaxUOK4qcRpA/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — BrowserStack capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "browserstack-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_QfMSmaBJhbb00bXFG15oIucKmIgIGaxUOK4qcRpA/mcp"
    }
  }
}
Copy into chat04

  • Get the logs for BrowserStack session 12345 and tell me why it failed.

  • Look at the logs from session 98765 and find the stack trace.

  • Pull the execution logs for my latest failed session and summarize the errors.

  • Check session 55443 to see if the failure was a timeout or a logic error.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect BrowserStack once, then ask.

One connection links your BrowserStack account to your agent. Your credentials stay encrypted, and you can start debugging immediately.

Connect BrowserStack to your AI

FAQ

How this task behaves.

  • 01

    Can the AI fix the code in my repository automatically?

    No. The AI identifies the error in the logs, but you must manually apply the fix to your code.

  • 02

    Can the AI start a new test session in BrowserStack?

    No. This MCP is designed to read existing session data and logs, not to trigger new test runs.

  • 03

    Does the AI see my entire BrowserStack account?

    The AI only accesses the specific session data you request through the tools.

  • 04

    Can the AI download the video recording of the session?

    No. The current tools only allow the AI to access text-based execution logs.

  • 05

    Do I need to keep the BrowserStack dashboard open?

    No. Once the MCP is connected, your agent pulls the data directly.

  • More questions about BrowserStack? The Connector page answers them. See everything the BrowserStack Connector can do

Connect BrowserStack to Claude, Cursor, ChatGPT & more