Bugpilot

What is Bugpilot?

Bugpilot is a Chrome extension that records browser debugging context and turns it into structured Markdown for AI assistants. It captures console output, network requests, DOM state, and the sequence of user actions that led to a bug, so developers can paste a single report into tools like Claude or ChatGPT.

The product is designed to reduce the back-and-forth that usually happens when an AI lacks enough reproduction detail. Instead of manually collecting screenshots, console logs, and network information, a user records a session, reproduces the issue, stops the capture, and exports the result in a format the AI can read.

Key Features

• Records console events, including errors, warnings, logs, timestamps, and stack traces, so debugging context is captured in one place.
• Captures network requests with URLs, methods, status codes, and response previews to document what the app was doing when the bug occurred.
• Stores DOM context, including the outer HTML at capture time and the element that was clicked, which helps explain the UI state behind the issue.
• Logs user actions such as clicks, form input, and navigation, providing a reproduction trail instead of relying on a written summary.
• Offers optional screenshots as WebP frames at recording start, on console errors, and at stop, with timeline metadata linking each frame to the triggering event.
• Applies automatic redaction for passwords, email addresses, credit card numbers, JWTs, API keys, authorization headers, and URL tokens before data is stored.
• Provides export formats beyond Markdown in the Pro tier, including Claude XML, OpenAI JSON, single-file HTML, and Markdown plus JSON ZIP bundles.
• Runs locally in the browser with no accounts, servers, telemetry, tracking, or analytics, so captures remain on the user’s machine until exported.

How to Use Bugpilot

Install the Chrome extension, then open the popup when you hit a bug in a web app. Click Record, reproduce the issue, then click Stop and copy the capture as Markdown or another supported format.

Paste the output into an AI coding assistant such as Claude or ChatGPT, or save it as a report for later review. If needed, enable screenshots or use a Pro export format before copying or downloading.

Use Cases

• Filing a bug report for an AI assistant: a developer records the issue once and pastes the structured output into Claude or ChatGPT instead of assembling logs manually.
• Sharing a reproduction package with a teammate: the captured user actions, console output, and network requests make it easier for another developer to follow the same steps.
• Investigating a UI interaction bug: DOM state and click-target capture help explain what element was actually involved when the problem occurred.
• Debugging API or request failures: the network timeline shows request details and status codes alongside the user action that triggered the failure.
• Capturing framework state for deeper debugging: the Pro tier includes React component tree, props, and hook state for cases where console logs alone are not enough.

FAQ

Does Bugpilot run in the browser or on a server? It runs locally in the browser. The source says there are no servers, telemetry, tracking, or analytics.

What does Bugpilot export? The free version exports Markdown and can download a ZIP with screenshots. The Pro tier adds several AI-oriented formats, including Claude XML, OpenAI JSON, HTML, and combined Markdown/JSON ZIP output.

Is redaction optional? No. The source says redaction is always on and cannot be disabled.

Does Bugpilot require an account? The source says there are no accounts.

Which assistants does it target? It is positioned for Claude, ChatGPT, Cursor, and other AI coding tools that can read Markdown or the Pro export formats.

Alternatives

• Manual DevTools capture: screenshots, copied console output, and pasted network details can be assembled by hand, but the workflow is slower and less structured.
• Generic bug-reporting tools: these often focus on collecting screenshots or form-based reports rather than packaging console, network, DOM, and action history for AI use.
• Session replay or error monitoring tools: these can help with debugging, but they usually center on observability or playback rather than a copy-paste report for an assistant.
• Plain markdown notes: a developer can write reproduction steps and paste logs manually, but that approach does not automatically capture browser state or redact sensitive values.