Blocks
4 blocks for a marketplace page
A marketplace page has to make demand and supply meet within three clicks. Five categories, the card versus table choice, the no-results search faults and the real counts.
Key points
- Cards for items compared by eye, tables for numeric attributes.
- Three search blocks out of 1.556: no faceted filter in the catalogue.
- The “no results” state is part of search, not a detail.
A marketplace page has to make demand and supply meet within three clicks. The catalogue indexes 3 search blocks, 13 table, 4 gallery, 10 profile and 4 checkout, read on 2026-11-01: five categories for a search and result page.
The counts come from the public category pages, measured on the same day. The number that weighs most is search: 3 blocks out of a 1.556-record catalogue. The part that decides the product is the one no registry publishes as a copyable file.
What the page has to do
A marketplace does not sell an item: it sells the certainty that the right item exists and that its seller answers. The results page has to show availability, total price with shipping and seller reputation on the same row. The detail page repeats those three facts, because most doubts start from the comparison between the two.
- Total price, shipping included, already on the result row.
- A filter that survives refresh and shared links.
- A “no results” state with three clickable alternatives.
- A seller page with reply times and returns, not stars alone.
The order of the blocks
| Section | Role on the page | Blocks available |
|---|---|---|
| search | find an item or a service | 3 |
| table | compare on attributes | 13 |
| profile | decide who to trust | 10 |
| checkout | pay in a single flow | 4 |
Search, table, profile, checkout, footer. Search opens the page because it is the task; the result follows at once and the seller profile is reached from the result, not the other way round. Checkout is the last step and gets designed as its own flow: on a marketplace it may hold several sellers in one order, which no catalogue block anticipates.
The blocks to start from
The tables list block, registry, files and category. The 7-file topbar block is classified as search by the registry: it holds a box and a dropdown, not a results page. The 24-file data table is the densest choice and deserves a read before install, because it brings a whole table library with it.
Search blocks
| Block | Registry | File | Category |
|---|---|---|---|
| Topbar 06 - Search Focused Topbar | Shadcn Space | 7 | topbar |
| Global search | flx | 1 | global |
Table blocks
| Block | Registry | File | Category |
|---|---|---|---|
| Data Table Core | niko-table | 24 | data |
| Data Table | niko-table | 16 | data |
Profile blocks
| Block | Registry | File | Category |
|---|---|---|---|
| Profile Form - 1 | shadcn-ui-clerk-components | 3 | profile |
| Forms 01 - Edit Profile Form | Shadcn Space | 2 | forms |
Checkout blocks
| Block | Registry | File | Category |
|---|---|---|---|
| Checkout 01 - Tabbed Billing Checkout | Shadcn Space | 2 | checkout |
| Checkout 01 | commercn | 1 | checkout |
What breaks in production
The costliest fault is search without an empty state: zero results turns into a blank page and the visitor thinks the site is broken. The second is the price changing between result and checkout when shipping is added late. The third is the table whose rows grow taller the moment a long description arrives.
- No empty state: a search with no results looks like an error.
- Price without shipping in the list: the total changes at checkout.
- A table losing its header after two scrolls across 50 rows.
- Filters kept in local state: lost on refresh and in links.
curl "https://ui.eneawork.it/api/blocks/category/table?limit=5"The response lists the available tables with source and files: read it to see which target financial data and which target generic lists. A marketplace needs the second family, with cells that hold text and status together.
A limit to declare
The catalogue has no faceted filters: the 3 search blocks are fields with dropdowns and none ships per-filter counts. On a marketplace past 200 listings the faceted filter is the component that decides conversion, and it gets built from scratch with its own query.
In the catalogue: data table, price block and all blocks. The guides: accessibility checklist, measured MUI alternatives and Radix, Base UI and React Aria.
Frequently asked questions
Cards or a table for marketplace results?
Cards when photos drive the choice, a table when the visitor compares price, distance and rating. The table shows more rows per screen and allows sorting; the card holds the image, which for used goods matters more than any attribute.
Why does the catalogue hold only 3 search blocks?
Because useful search needs an index, ranking and a query: no registry ships that as a file. The 3 blocks are the box and the dropdown, the visible part. The engine stays backend work, and on a marketplace it is the part that costs most.
How do I show seller reputation without misleading?
With transaction count, average reply time and return rate, not the star average alone. A 4.9 average across three reviews says nothing, and next to a high average you need the count to judge how far to trust it.
Should checkout handle several sellers in one order?
If the marketplace allows it, yes, and then the 4 catalogue checkout blocks are not enough: each assumes a single cart. Splitting per seller touches totals, shipping and returns, so decide it before picking the block.
How many listings before filters become necessary?
Around 50, when results pass two screens. Below that threshold, sorting and text search are enough; above it, the faceted filter becomes the main feature and the catalogue offers none ready made.