Payload CMS Plugins We Actually Install (and Skip)
A Payload plugin isn’t free, it grows your admin bundle, adds collections to migrate, and can break on the next major version. We score the ones we install, skip, and replace with about 20 lines of custom code.

Originally published at techpotions.com.
Every Payload CMS plugin roundup reads like a feature list. Ours is a bill. A plugin is not free, it grows your admin bundle, it adds collections you have to migrate, and it becomes one more thing that can break when Payload ships a major version. So we score every plugin we consider on three costs: how much it adds to the admin bundle, how much migration burden it carries, and how badly an upgrade could hurt. The question we actually ask is "what will I have to fix when Payload 3.4 drops?"
Here are the ones that earn their place, the ones we only reach for in narrow cases, and the two we tore out and replaced with about twenty lines of custom code after a live-site meltdown.
We measured every bundle number the same way: a fresh Payload 3.0.3 and Next.js install, built with ANALYZE=true next build and @next/bundle-analyzer.
The Payload CMS plugins we actually install, and the ones we scrap for 20 lines of code
SEO plugin: always install, but don't blindly trust the defaults
The official SEO plugin goes on every public-facing site we build. It is light, about 18 KB gzipped, and under the hood it is a meta field group plus a tabbedUI helper, nothing exotic. The one gotcha is the defaults: uploadsCollection is undefined out of the box, so you have to point it at your media slug by hand or your Open Graph images never resolve. The real limit is that you get little control over JSON-LD beyond Open Graph, if you need rich structured data you will be adding it yourself. Worth it anyway.
Search plugin: use for simple blogs, skip for anything custom
This is the biggest admin-bundle liability we have shipped. The official search plugin adds around 210 KB gzipped, spins up its own search collection, and after v3.1 it needed a manual reindex that was effectively undocumented, we found the fix by reading the plugin's source. For a simple blog where you want title-and-excerpt search, it is fine. For anything custom it is the wrong tool. On the Work Well mats finder we pulled it out and replaced it with a small custom endpoint. Payload's own guide to the search plugin is worth reading before you commit to it.
Import/export plugin: great for one-off migrations, dangerous in production
The official import/export plugin is the right call for a one-time migration, then it should come out. We used it to bring 1,200 SKUs across from a legacy WooCommerce store, it is about 40 KB gzipped and adds no new collection. The catch is that it drops a bulk-export button onto every collection's list view, which is not something you want sitting in a production admin for editors to hit. For ongoing reporting we swapped it for a custom CSV export endpoint, twelve lines, see Payload's import-export docs for what it does before you lean on it.
What we replaced with twenty lines of custom code: the quote builder and mat finder
Two plugins never made it to production. A community form builder looked like a shortcut for a quote builder, but it added 60 to 120 KB, created extra collections, and broke entirely on Payload 3.0. We replaced it with a Custom Component, about fifteen lines, no new collections. A PDF exporter plugin went the same way, rejected and rebuilt with @react-pdf/renderer in a Next.js API route in under forty lines.
The mat finder is the clearest example. It is a three-question flow, questions → answers → products, that we first tried to force through the search plugin. The custom version is about twenty lines of logic over a three-line local API join on floorMaterial, oilResistance, and shape, and it has no bundle cost and nothing to break on upgrade. The admin panel is rich enough that a Custom Component often matches a plugin's UI with none of the production surprises.
The Payload CMS plugins that make the grade
Here is the whole shortlist at a glance.
| Plugin | Admin bundle impact | Migration burden | Upgrade risk | Verdict |
|---|---|---|---|---|
| Official SEO plugin | ~18 KB gzipped | Low, just a field group | Very low | Install on every project |
| Official search plugin | ~210 KB gzipped | Medium, separate search collection, index rebuild after v3.1 | Medium | Only for simple blog search |
| Official import/export plugin | ~40 KB gzipped | Low, no new collection | Low | Use for migration, then remove |
| Any community form builder | 60 to 120 KB gzipped | High, extra collections and custom field types | High, broke entirely on Payload 3.0 | Skip, write 15 lines of custom component code |
Bundle numbers measured on a fresh Payload 3.0.3 and Next.js install with ANALYZE=true next build.
Authentication, media transforms, and i18n are usually better handled with Payload's built-in hooks and custom fields than with a third-party plugin. We only reach outside for something like multi-tenant auth or direct Cloudflare R2 uploads.
FAQ
Which Payload CMS plugins should I install first?
Start with the official SEO plugin on any public site, it is cheap and you will want it. Add the import/export plugin only when you have a migration to run, and remove it afterwards. Hold off on the search plugin unless your search really is just blog-simple.
How do I evaluate a Payload plugin's impact on admin performance?
Build with ANALYZE=true next build using @next/bundle-analyzer and read the admin bundle before and after you add the plugin. As a rough line, under 30 KB gzipped is healthy. Also count the collections it creates, each one is migration and upgrade surface, and ask what breaks when the next major Payload version lands.
Can I build custom functionality without plugins in Payload?
Often, yes, and it is frequently the better call. A Custom Component plus a few hooks can match a plugin's UI in a dozen or two lines, with no new collections and nothing extra to fix on upgrade. We have replaced form builders, PDF exporters, and search with small custom code more than once, including on the Work Well mats finder.
We build production Payload and Next.js sites this way, lean on plugins where they pay off, custom where they don't. See our web work or Next.js development, or tell us what you're building and we'll reply within 24h.



