Why @next/bundle-analyzer wasn’t enough for our team (So we built an automated PR bot)

Description and use

An automated GitHub bot checks bundle size in PRs and warns of regressions, eliminating manual @next/bundle-analyzer analysis.

Problem

Manual bundle analysis with @next/bundle-analyzer is too laborious for fast‑moving teams, leading to unnoticed bundle regressions.

Outcome

Automated PR bot reduced bundle regression incidents by eliminating manual checks and providing instant feedback within PRs.

Workflow

• 1. Vercel preview deployment triggers webhook on success.
• 2. Webhook payload is intercepted by NextPerf.
• 3. NextPerf retrieves production baseline JS size from edge storage.
• 4. It calculates PR JS size and difference.
• 5. If difference exceeds threshold defined in .nextperf.json, NextPerf posts a markdown table with Base Branch, This PR, and Difference metrics and a warning message.
• 6. Developers review the comment and address regressions before merging.

Evidence

To check your bundle size, a developer has to: 1. Remember to run a local build with an env flag (ANALYZE=true npm run build). 2. Wait for Webpack/Turbopack to finish compiling. 3. Open a massive, interactive HTML treemap in their browser. 4. Go hunting through nested colored boxes to figure out which file grew.

Source

We wanted it to live right where code reviews happen: GitHub Pull Requests.

Source

Now, whenever someone opens a PR on our repo, a clean little summary drops in automatically: ...

Source

Tools like @next/bundle-analyzer are fantastic for deep-dive optimization sessions when you're actively trying to trim fat off your app. But for everyday development, automation is the only way to prevent regressions permanently.

Source