OPENAI BUILD WEEK PROJECT

boring Bugshot

One screenshot in. One bug report out.

Drop in what looks wrong. Get a structured, editable ticket your team can actually use.

GPT-5.6 visionStructured OutputsServer-validated schemaIssue-ready exports
+

Drop or paste your screenshot

Press Ctrl+V anywhere · or click to browse · max 4 MB

Before you analyze: Your screenshot, optional context, report type, and output language are sent to OpenAI through our server to create the report. Bugshot does not add the screenshot to your saved browser workspace. Do not upload confidential or personal data unless you are authorized to share it. Privacy details

Built with Codex for OpenAI Build Week.

GPT-5.6 reads the screenshot. Structured Outputs turn the result into an editable, server-validated report — a focused workflow instead of a chat wrapper.

FROM VISUAL FIND TO USEFUL TICKET

You can see the problem.
Now make it actionable.

A screenshot often explains what is wrong faster than a meeting. But screenshots alone leave developers guessing. boring Bugshot closes the gap with one focused, structured report.

01

Capture it

Take a screenshot and paste it straight from your clipboard. Add context only when it helps.

02

Structure it

Vision input and schema-validated output turn the visible issue into a consistent bug report.

03

Use it

Review the result, fix the wording, save it locally, or export it as PDF or ticket-ready Markdown.

WHAT YOU GET

Enough structure.
No dashboard circus.

Clear titleA short summary people can scan.

Category & priorityUseful triage without a complex QA system.

Problem & impactWhat is visible and why it matters.

Suggested fixA practical starting point, ready for review.

Ticket-ready textObserved result, expected result, and acceptance criteria.

BUILT FOR THE HANDOFF

For people who find bugs.
And people who have to fix them.

Developers

Designers

Freelancers

Agencies

Small product teams

Client projects

TECHNICAL IMPLEMENTATION

Vision in.
Structured data out.

Bugshot sends the screenshot and optional context to GPT-5.6 through the Responses API. Structured Outputs are parsed into a Zod schema on the server before the interface can use them.

The model is the product core. The interface turns it into a dependable handoff.

THIS BROWSER

Saved reports

No saved reports yet. Your first useful ticket will land here.