Windowed Grid Primitive
Keep very large tables and grids responsive by rendering only the cells currently on screen.
What it adds
A two-dimensional windowing layer that renders visible rows and columns while preserving normal table semantics and navigation.
What your agent is told to do
5
What your agent is told to do
5-
1
Identify the surfaces that genuinely need windowing — thousands of rows, or wide column sets rendered at once — and leave the rest alone. A list of forty rows is slower to window than to render.
-
2
Support rows and columns of differing sizes by estimating first and correcting once measured, and cache the measurements so scrolling back does not re-measure and jump.
-
3
Keep keyboard focus stable when the focused cell scrolls out of the window. Either keep that cell mounted or move focus deliberately to a known position; never let it fall back to the document body.
-
4
Restore scroll position by the identity of the item that was at the top, not by a pixel offset, because a pixel offset points somewhere else as soon as sizes or filters change.
-
5
Publish the true row and column counts and each cell's real position to assistive technology, so the grid announces a cell as one of ninety thousand rather than one of the forty currently rendered. Memoisation of the cell renderer belongs to the Render Stability Guardrails brief; this brief owns which cells exist, not how cheaply each one re-renders.
Edge cases it handles
7
Edge cases it handles
7- Variable row heights and column widths must be handled, not assumed away. A grid built on a fixed row height will misplace every item below the first row that wraps to two lines.
- Focus must survive scrolling and data updates. A user tabbing through cells who is thrown to the top of the document has lost their place and, with a screen reader, their context entirely.
- Scroll restoration has to land on the same record after a reload or a return from a detail view. Restoring to a stored pixel offset after the data changed puts the user in an arbitrary part of the grid.
- Assistive technology must be given the grid's real structure. Without explicit counts and positions it reports only the rendered window, which makes the grid unnavigable.
- Find-in-page, select-all, and copy only reach rendered cells. Provide an explicit export or copy path for the full dataset and say so, rather than letting users think they copied everything.
- Resizing the window, resizing a column, or toggling density mid-scroll invalidates cached measurements and must trigger a recalculation that keeps the current top item in place.
- Sticky headers and frozen columns must share the same scroll arithmetic as the body, or they drift apart by a row at speed.
Definition of done
9
Definition of done
9- Only cells within and just beyond the viewport are rendered, at any scroll position.
- Rows and columns of differing sizes render in the correct positions with no drift or jumping.
- Keyboard focus remains on a known cell through scrolling, filtering, and data updates.
- Scroll position restores to the same record after a reload or a return to the view.
- Assistive technology reports the full row and column counts and each cell's true position.
- Sticky headers and frozen columns stay aligned with the body during fast scrolling.
- There is an explicit path to copy or export the full dataset rather than only the rendered window.
- The feature matches the existing design system.
- No existing functionality is broken.
Related features
Font Subsetting
Font Subsetting
Load only the glyphs a page needs, without breaking the alphabets it did not expect.
What it does
Per-script font subsets with declared character ranges and a fallback order for characters no subset covers.
How it works
- 1 Inventory the weights, styles, and families the app genuinely renders. Most projects ship two or three weights they never use, and dropping those is a larger win than any subsetting.
- 2 Split each family by writing system and declare the character ranges each file covers, so a page that renders only Latin text never fetches the Cyrillic or Greek file.
- 3 Choose subsets from the locales the product supports and the languages its users actually write in, not from the words currently on the page.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/font-subsetting
Layout Shift Guard
Layout Shift Guard
Stop content jumping as a page loads, and catch it when it starts happening again.
What it does
Reserved space for everything that arrives late, plus measurement that reports unexpected movement on the app's main screens.
How it works
- 1 Walk the app's main screens with the network throttled and note everything that moves after first paint. Images, avatars, embeds, banners, ad slots, and asynchronously loaded panels are the usual sources.
- 2 Give every image, video, iframe, and embed explicit dimensions or a fixed aspect ratio so its box exists before its content arrives, including the ones whose size comes from data.
- 3 Control the swap from fallback to web font so the two have compatible metrics, and accept a slightly imperfect fallback over a page that reflows once the real font lands.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/layout-shift-guard
Adaptive Media Source Component
Adaptive Media Source Component
Serve each image and video at a size and format the viewer's device can actually use.
What it does
One media component that chooses source, format, and dimensions from the rendered box and the browser's declared support.
How it works
- 1 Find every place the app renders a raw image or video element and route them all through a single component, so source selection is decided in one place rather than per template.
- 2 Derive the candidate widths from the sizes the element is genuinely rendered at across the app's breakpoints, and describe those sizes to the browser accurately — an inaccurate size hint is worse than none.
- 3 Always render an intrinsic width and height or an aspect ratio, so the box occupies its final space before any bytes arrive.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/adaptive-media-source-component
How it works
-
1
Copy the link
Grab the Markdown instruction URL for this feature.
-
2
Give it to your AI
Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.
-
3
It inspects, then implements
Your agent reads your existing app first, then adds the feature to fit it.
Works with your stack
These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.
Need it tighter than that? Customize the feature and tell it exactly what you're running.