AI Fix

Overview

AI Fix lets QA engineers and product managers fix simple bugs — like typos, wrong colors, outdated copy, and minor logic issues — directly from the bug report, without involving developers. The AI agent reads the bug context, analyzes the connected GitHub repository, proposes the exact code change, and creates a pull request.

This is ideal for small fixes that don't require deep engineering context — the kind of changes that often sit in a backlog for days because they're too small to prioritize but still need to be done.

How it works

1. Open a track and go to the AI Fix tab

Navigate to any track on the VibeCheck platform. Click the AI Fix tab in the track detail view.

2. Select a GitHub repository

Choose the repository where the bug exists from the dropdown. You must have GitHub connected in Settings with repo scope for this to work.

3. Click "Fix this bug"

The AI agent starts analyzing your bug report. It reads:

• Console errors — JavaScript errors and warnings captured during recording
• Network failures — failed API requests with response data
• User actions — clicks, inputs, and navigations that led to the bug
• Page URL — the page where the bug was observed

The agent streams its progress in real time so you can see what it's doing — reading files, searching code, and narrowing down the fix.

4. Review proposed changes

The agent proposes code changes with inline diff previews. You can see exactly what will change in each file before anything is committed.

5. Create a Pull Request

Click Create Pull Request to push the changes. The agent creates a new branch, commits the files, and opens a PR on GitHub — ready for developer review and merge.

What AI Fix can fix

AI Fix works best for small, straightforward changes:

• Typos — misspelled words in UI text
• Copy updates — changing button labels, headings, descriptions
• Color changes — updating hex codes, Tailwind classes, CSS values
• Minor logic fixes — simple conditional changes, off-by-one errors
• Configuration tweaks — updating constants, feature flags, environment values

What AI Fix is not for

AI Fix is not a replacement for developers. It won't handle:

• Complex refactors or architectural changes
• Multi-file feature implementations
• Performance optimizations
• Security fixes
• Database schema changes

For these, use the bug report as context and assign it to a developer.

Requirements

• GitHub connected — go to Settings > Integrations on the platform and connect your GitHub account
• Repo scope — the GitHub OAuth token must have repo scope to read code and create PRs
• Repository access — you must have write access to the repository you want to fix

Usage limits

Usage is counted only when the agent successfully proposes changes. Failed attempts (e.g., the agent couldn't find the relevant code) don't count toward your limit.

Tips

• Write clear bug descriptions — the more context in the bug report, the better the AI fix. Include what's wrong and what it should be.
• Use tab recording — tab recordings capture console errors and network data, which gives the agent more context than screenshots alone.
• Review before merging — always review the proposed diff. AI Fix creates a PR so developers can review before merging.
• Start with obvious fixes — typos and copy changes have the highest success rate. Try these first to get a feel for the feature.