Track AI-generated changes in GitHub using commit history, blame view, and bot author labels. Lovable commits appear under the lovable-dev bot account, V0 commits come from V0's bot on separate branches, and Cursor marks AI edits with a sparkle icon locally. Use GitHub's blame view to trace any line of code back to the exact commit and author that created it.
Why Tracking AI Changes Matters
When you use AI tools to build your app, dozens or hundreds of commits pile up in your GitHub repository. Some come from AI, some from you, and some from collaborators. Without a way to tell them apart, debugging becomes guesswork — you do not know which change introduced a bug, who approved it, or whether a human ever reviewed it.
GitHub provides three key features for tracking changes:
1. Commit history — a chronological list of every change, showing who made it and when. Each commit has an author, a message, and a diff (the exact lines added or removed).
2. Blame view — a per-file view that shows which commit last modified each line. This lets you trace any line of code back to the exact commit and author that created it.
3. Author filtering — you can filter the commit history to show only commits from a specific author, like the lovable-dev bot.
Different AI tools leave different fingerprints on GitHub. Lovable uses a bot account (lovable-dev) so its commits are instantly recognizable. V0 creates commits on separate branches with its own bot. Cursor and Replit commit under your personal account, making AI changes harder to distinguish — but Cursor marks AI edits with a sparkle icon in its local editor.
Building a habit of reviewing AI commits regularly is the single most important practice for maintaining code quality in AI-built projects.
Prerequisites
- A GitHub repository with commits from at least one AI tool
- Access to the repository on github.com
- Understanding of what commits are (saved snapshots of code)
Step-by-step guide
Open the commit history for your repository
Open the commit history for your repository
Navigate to your repository on github.com. On the Code tab, look near the top of the file list — you will see a clock icon with text like '142 commits.' Click that text to open the full commit history. This page lists every commit in reverse chronological order (newest first). Each entry shows the commit message, the author avatar and name, and the date. Look for commits authored by lovable-dev, v0-bot, or other AI accounts.
Expected result: You see a chronological list of all commits, with author names and avatars visible for each one.
Filter commits by AI bot author
Filter commits by AI bot author
On the commit history page, look for the search bar at the top. You can type an author's name to filter. For Lovable commits, look for entries with the lovable-dev bot avatar (a purple/blue icon). You can also add '?author=lovable-dev' to the end of the URL in your browser to filter directly. For V0 commits, check the Pull requests tab instead — V0 creates branches and PRs rather than committing directly to main.
Expected result: The commit history is filtered to show only commits from the AI bot, making it easy to review all AI-generated changes.
Use blame view to trace a specific line of code
Use blame view to trace a specific line of code
Navigate to any file in your repository by clicking through the folder structure on the Code tab. Once you open a file, click the Blame button in the top-right area of the file viewer (it is next to the Raw and Edit buttons). Blame view shows each line of code with the commit that last modified it, the author, and the date. This is invaluable when you find a bug — you can see exactly which commit introduced the problematic line and whether it was an AI or human change.
Expected result: Each line of the file shows the commit hash, author, date, and message of the last change, letting you trace any line back to its origin.
Review a specific AI commit in detail
Review a specific AI commit in detail
From the commit history, click on any commit to see its full diff. Green-highlighted lines are additions, red-highlighted lines are deletions. The top of the page shows how many files were changed and the total lines added/removed. For large AI commits (which are common — AI tools often modify many files at once), click individual file names in the diff to jump to specific changes. Pay special attention to files related to authentication, database queries, and API keys.
Expected result: You can see the exact code added, removed, or modified in any AI-generated commit.
Set up a weekly AI commit review routine
Set up a weekly AI commit review routine
Bookmark your repository's commit history page. Once a week, open it and scan for AI-authored commits you have not reviewed. For each one, click to view the diff and check for unexpected changes — deleted files, modified environment configurations, or added dependencies. If you spot an issue, use the Revert button. For repositories with heavy AI activity, consider adding a CODEOWNERS file (Settings, then search for CODEOWNERS in the repo) to require review notifications.
Expected result: You have a repeatable weekly process for auditing AI-generated changes in your GitHub repository.
Complete working example
1# Weekly report of AI-authored commits.2# Posts a summary as a GitHub Issue every Monday.34name: AI Commit Report56on:7 schedule:8 - cron: '0 9 * * 1' # Every Monday at 9 AM UTC910jobs:11 report:12 runs-on: ubuntu-latest13 steps:14 - uses: actions/checkout@v415 with:16 fetch-depth: 01718 - name: Count AI commits this week19 id: count20 run: |21 SINCE=$(date -d '7 days ago' --iso-8601)22 LOVABLE=$(git log --since="$SINCE" --author="lovable" --oneline | wc -l)23 echo "lovable_count=$LOVABLE" >> $GITHUB_OUTPUT24 echo "Found $LOVABLE Lovable commits this week"2526 - name: Create issue27 uses: peter-evans/create-issue-from-file@v528 if: steps.count.outputs.lovable_count > 029 with:30 title: "Weekly AI Commit Report"31 content: |32 Lovable commits this week: ${{ steps.count.outputs.lovable_count }}33 Review at: ${{ github.server_url }}/${{ github.repository }}/commitsCommon mistakes when tracking Changes Made by GPT and AI Tools in GitHub
Why it's a problem: Assuming all AI commits are safe because the tool generated them
How to avoid: AI tools can introduce bugs, delete files, or expose secrets. Always review the diff of AI commits, especially for authentication and payment code.
Why it's a problem: Not knowing that Cursor and Replit commits appear under your personal account
How to avoid: Unlike Lovable's bot, Cursor and Replit use your GitHub account for commits. Add '[AI]' or '[Cursor]' to commit messages manually to identify AI-assisted changes.
Why it's a problem: Ignoring large AI commits because they are overwhelming to review
How to avoid: Focus on critical files first: authentication, database, payments, environment config. You can review these files individually in the diff view.
Why it's a problem: Not using blame view when debugging
How to avoid: Blame view is the fastest way to find which commit changed a buggy line. Click Blame on any file to trace every line to its source.
Best practices
- Check the commit history weekly and review all AI-authored commits you have not seen.
- Use blame view to trace bugs back to the specific AI commit that introduced them.
- Filter commits by author to quickly see all AI-generated changes.
- Add '[AI]' tags to manual commit messages when using Cursor or Replit to distinguish AI-assisted work.
- Set up a CODEOWNERS file to get notified when AI tools modify critical files.
- Use the Revert button on GitHub to safely undo problematic AI commits.
- Prioritize reviewing commits that touch authentication, payments, database, and API key files.
- Bookmark filtered commit history URLs for quick access to AI bot activity.
Still stuck?
Copy one of these prompts to get a personalized, step-by-step explanation.
I have a GitHub repository with 200+ commits from Lovable's bot and 50 manual commits. I found a bug in my authentication flow. Walk me through how to use GitHub's blame view and commit history to trace the bug back to the specific commit that introduced it.
Frequently asked questions
How do I know if Lovable's bot made a specific commit?
Lovable commits are authored by the 'lovable-dev' bot account. You can see this in the commit history — the author avatar and name will show lovable-dev instead of a personal account.
Can I see which AI tool made a change if I use Cursor?
Cursor commits appear under your personal GitHub account, not a bot. Cursor does mark AI-edited lines with a sparkle icon in its editor, but this information does not transfer to GitHub. Add '[Cursor AI]' to your commit messages manually.
What is GitHub blame view?
Blame view shows which commit last modified each line of a file. Open any file on GitHub and click the Blame button to see the author, date, and commit message for every line. It is the fastest way to trace a bug to its source.
Can I revert an AI commit without affecting other changes?
Yes. On GitHub, open the specific commit page and click the Revert button. This creates a new commit that undoes only that change. Other commits before and after remain untouched.
How can RapidDev help with tracking AI changes?
RapidDev's engineering team can set up automated GitHub workflows that flag AI-generated commits for review, create weekly audit reports, and configure CODEOWNERS rules for your repositories.
Does GitHub show which lines were AI-generated within a commit?
GitHub shows which lines were changed in a commit (the diff view), but it does not distinguish between lines the AI wrote and lines a human wrote within the same commit. That distinction only exists in tools like Cursor that mark AI edits locally.
Talk to an Expert
Our team has built 600+ apps. Get personalized help with your project.
Book a free consultation