AI-Powered Playwright Test Generation

Generate, Execute & Self-Heal Tests Directly From Jira Tickets

QBot transforms Jira requirements into production-ready Playwright automation using AI, real DOM analysis, code diffs from Bitbucket, automated linting, execution, and self-healing repair loops.

Watch It Work

An animated walkthrough of the QBot pipeline — from Jira ticket to self-healed tests.

QBot — Pipeline Runner
🎫

Fetch Jira Ticket

Connecting to Jira and retrieving requirements.

End-to-End AI Testing Pipeline

From Jira ticket to executed Playwright tests in minutes.

01

Fetch Jira Ticket

Connect to Jira Cloud or Server/DC and retrieve requirements, acceptance criteria and comments.

02

Crawl Application

Capture real DOM structure, forms, selectors, links, dropdowns and ARIA snapshots.

03

Fetch Code Changes

Locate linked Bitbucket commits and extract relevant diffs for implementation awareness.

04

Generate Tests

AI creates Playwright tests using ticket requirements, DOM context and code changes.

05

Lint & Auto-Fix

Detect hallucinated selectors and invalid patterns before execution.

06

Execute

Run tests in Chrome using saved authentication state and retry flaky failures.

07

AI Self-Healing

Failures are analyzed and repaired automatically using real error output.

08

Results

View pass/fail status, human-readable explanations and raw pytest output.

Why Teams Use QBot

Built for modern QA, SDET and engineering teams.

Real DOM Understanding

No blind AI generation. QBot crawls actual pages and uses real selectors.

Bitbucket-Aware Testing

Code diffs provide implementation context for higher quality test coverage.

ARIA Accessibility Snapshots

Generate robust get_by_role() locators and accessibility-aware tests.

Self-Healing Automation

Failed tests are repaired automatically using actual execution errors.

Team Profiles

Customize testing style, selector conventions, technology stack rules and product glossary.

Smart Context Management

Token budgeting, context trimming and intelligent prompt optimization.

Setup & Configuration

Get QBot running in minutes.

Prerequisites

  • Python 3.12 or higher
  • Google Chrome browser
  • GitHub Copilot subscription (org or individual)

No external API keys needed — QBot uses the same Copilot API that powers VS Code Copilot Chat.

Quick Install

  • From Source:
  • pip install -r requirements.txt
  • playwright install chromium
  • python main.py

GitHub Copilot Authorization

  • Open QBot Settings (gear icon)
  • Click Authorize Copilot
  • Browser opens to github.com/login/device
  • Enter the device code shown in QBot
  • OAuth token is saved and auto-refreshes

Token persists across app restarts. No manual refresh required.

Connect Your Services

  • Jira: Cloud (atlassian.net) or Server/DC with API token or PAT
  • Bitbucket: Cloud API token for linked commit context
  • Target URLs: Add your app URLs in Settings for quick access

Check "Remember credentials" to auto-fill on next launch.

Available AI Models

Select the model that best fits your team's needs.

Recommended
Claude Sonnet 4.5
200K context
Most Capable
Claude Opus 4.5
200K context
Fastest
Claude Haiku 4.5
200K context
GPT-5.5
272K context
Gemini 3.1 Pro
200K context
GPT-4o
128K context

Switch models anytime in Settings → Team Profile. Each team can have different model preferences.

Built With Security In Mind

Your credentials remain local and are never sent to AI models.

What Gets Sent To AI

  • Jira ticket content
  • DOM snapshots
  • ARIA snapshots
  • Bitbucket code diffs

What Never Leaves Your Machine

  • GitHub OAuth tokens
  • Jira credentials
  • Bitbucket API tokens
  • Browser session cookies