PixelOkay
On-device analysis
Tools

PIXEL·OKAY / DEVELOPER STUDIO

Bug Reproduction Studio

Turn scattered errors, console output, and network traces into an actionable bug report. Correlate failed requests, surface possible causes, and document a clear sequence to reproduce—without uploading project data.

Private by designHAR + console importEvidence-linked suggestionsMarkdown / JSON / Issue template
Add diagnostics or choose the sample case to explore the workflow.
01

Import evidence

All processing stays in this tab.

Drop log or HAR files hereor choose files from your device.har or .json (HAR / saved session): 16 MB · .log, .txt: 650k characters
No imported files
Paste a network request trace (optional)

Prefer a HAR file. For text, use lines such as POST /api/drafts 500.

Privacy note

We strip common credentials, email addresses, and URL query values from parsed evidence. No redactor can catch everything; inspect the report before sharing.

02

Reproduction context

Document what actually happened.

Console events—
Network requests—
Failed requests—
Investigate—

Diagnostic workbench

Transparent, rule-based pattern matching. Hypotheses are not confirmed diagnoses or generative-AI output.

Your bug investigation starts here

Import a HAR file, paste a console error, or enter reproduction steps. Analyze to see linked failure signals and recommended checks.

From debugging evidence to a reproducible bug report

01 / Capture

Collect console and network evidence

Export a sanitized HAR from your browser's Network panel, paste JavaScript exceptions, and add steps you personally observed.

02 / Correlate

Investigate concrete signals

The local rules highlight HTTP failures, known exception classes, slow endpoints, and time/URL overlap. They cannot prove root cause.

03 / Share

File a report your team can act on

Review the reproduction sequence, expected and actual results, supporting requests, and caveats. Export an issue template, Markdown or JSON, or print to PDF.

Does this send my logs to an AI service?

No. Deterministic reports use browser-only rules and work without any model. An optional WebLLM model can be downloaded only after your consent to review a redacted report summary locally; no hosted inference API is used. AI suggestions are unconfirmed hypotheses.

What is a sanitized HAR?

A browser network export that removes some credentials and cookies. HAR files can still contain personal or confidential information, so this tool discards request and response bodies and aggressively limits what is included in reports.

Using Bug Reproduction Studio

Turn console logs, HAR traces and your reproduction notes into an evidence summary that helps a developer investigate a possible bug. Inspect the rule-based leads, fill the gaps manually and export a report for review.

  1. Import a HAR file, console text or the built-in example, then add the browser, steps and expected result.
  2. Review repeated errors, relevant requests and rule-based diagnostic hypotheses; edit the reproduction details and verify every lead against the source evidence.
  3. Export Markdown or JSON, save a portable session, or print the reviewed report.

Example: Import a HAR and a console log from a failed sign-in, add the exact steps and expected behavior, then review the repeated status/error clues before exporting an issue draft.

Limits and privacy: This is not autonomous reproduction, a confirmed root-cause detector or a language model. Redaction is best-effort; review the complete report and portable session before sharing. HAR request bodies, cookies and raw headers are discarded; each imported file is limited to 16 MiB. Optional local WebLLM suggestions are separate and require explicit setup; model/runtime behavior remains unverified.

Related tools

LOCAL AI SETUP

Set up AI on this device

First, a model needs to be loaded on your device

Bug Reproduction Studio uses local WebLLM inference. First-time setup may download hundreds of megabytes. No model download starts automatically. The ordinary report builder works without AI.

  1. 1. Check device
  2. 2. Download / cache
  3. 3. Set up assistant
  4. 4. Start review

Device & model resources

Measurements may be approximate
Device RAM / model memoryChecking device information…

Live available RAM and GPU memory are not exposed by most browsers.

Browser website storageChecking site storage…

Site-wide approximate origin quota, shared with other PixelOkay tools. This is not RAM.

ONE ACTION · ALL RECOGNIZED AI MODEL DOWNLOADS

Free AI model storage across PixelOkay

Scan this website origin for WebLLM and Transformers.js model caches, plus recognizable model files in other cache buckets. This includes caches from other PixelOkay tools, not only this tool. No site data or browser databases are cleared automatically.

Run a scan to inspect recognized model caches before optional cleanup.

Unknown IndexedDB, OPFS and browser-managed models remain untouched. Other AI tabs can hold models in use; close them before cleanup.

Having trouble downloading or loading AI?

Refresh storage/cache, retry, select a smaller compatible model, or inspect the recognized caches. Storage deletion is optional and does not fix network/GPU problems.

Clearing all website data can erase saved work across PixelOkay; this tool never performs that action.

Check compatibility or select Download & Load Model to begin.

Technical details and safe recovery

WebLLM 0.2.87 · selected model. No private evidence in diagnostics.

Do not clear all browser website data without exporting your projects and files. This tool does not access or remove them.

The model analyzes only the bounded, redacted report summary on your device. AI suggestions are not proof of reproduction or a confirmed root cause.
CROSS-TOOL MODEL CACHE CLEANUP

Delete recognized AI model downloads?

This may make other PixelOkay AI tools redownload shared model weights. Saved projects, imported files, and browser databases are not targeted. Unknown caches are skipped. Close other AI tabs first.