Tag Input
A field where typing a word and pressing Enter turns it into a removable chip.
What it adds
The input control for attaching tags to a record: type to filter, Enter to commit, Backspace to remove, chips for what is already there.
What your agent is told to do
5
What your agent is told to do
5-
1
Build the control only. The tag catalog itself — scoping, renaming, deleting, filtering lists by tag — is owned by Labels and Tags; extend that rather than building a second one.
-
2
Normalise on commit, not on every keystroke: trim whitespace, collapse runs of spaces, apply the app's case rule. Rewriting the value while someone is mid-word makes the field feel broken.
-
3
Support two modes: create-new-allowed and existing-only. In existing-only mode the field must say so rather than silently swallowing an unmatched entry.
-
4
Make the whole thing keyboard-drivable: arrow keys through suggestions, Enter to commit, Backspace on an empty input to target the last chip, Escape to close the suggestion list.
-
5
Do not re-specify listbox roles and announcements — Accessible Multi-Select owns those semantics. Reuse them.
Edge cases it handles
6
Edge cases it handles
6- Punctuation people mean to keep, like C++ or .NET, must survive normalisation.
- A tag that normalises to one already attached is a duplicate — reject it quietly and flash the existing chip rather than adding a second.
- Backspace must delete the last chip in one press only after the input is empty, and it should select before it destroys.
- Pasting a comma- or newline-separated list should create several tags at once, not one tag containing commas.
- A very long tag must not push the input off the row; wrap or truncate the chip, never the field.
- Removing a chip must return focus somewhere sensible, not to the top of the document.
Definition of done
8
Definition of done
8- Tags can be created, selected, and removed without touching the mouse.
- Normalisation happens on commit and never rewrites the value mid-typing.
- Duplicates after normalisation cannot be added.
- Existing-only mode refuses unknown values with a visible message.
- Pasting a delimited list produces separate tags.
- The catalog itself is read from Labels and Tags, not duplicated here.
- The feature matches the existing design system.
- No existing functionality is broken.
Related features
Voice Dictation
Voice Dictation
Put a microphone in text fields so people can speak their input instead of typing it.
What it does
A dictation control on the app's longer text inputs that transcribes speech into the field as the user talks.
How it works
- 1 Add the control to the fields where typing is genuinely tedious: long descriptions, notes, comments, and message bodies. A microphone on a postcode field is clutter.
- 2 Show interim results in the field as they arrive, visually distinct from committed text, so the speaker can see it is working and correct course before they finish.
- 3 Reuse the app's existing Voice Transcription for the server-side path so one recognition configuration and one set of language settings serve both dictation and recorded audio.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/voice-dictation
Help Text Pattern
Help Text Pattern
Put field guidance in one predictable place, phrased the same way every time.
What it does
A consistent treatment for the short explanatory text that sits with a field and its control.
How it works
- 1 Collect the guidance the app currently scatters across placeholders, tooltips, asterisked footnotes, and modal help, and move it into one slot beneath the field.
- 2 Keep guidance that is required to complete the field permanently visible. A format rule hidden behind a hover target is not available to a touch user or a keyboard user.
- 3 Write help as the rule the user needs before typing — the accepted format, the consequence of the choice — not as a restatement of the label.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/help-text-pattern
Form Section Pattern
Form Section Pattern
Break long forms into labelled sections people can scan instead of read.
What it does
A section pattern grouping related fields under a heading, a description, and its own actions.
How it works
- 1 Find the forms that have grown past a screenful — settings, profiles, checkout, configuration — and group their fields by the question each group answers, not by the table the columns come from.
- 2 Give each section a heading in the document's heading order and an optional one-line description, so screen reader and keyboard users can jump between sections.
- 3 Allow at most one level of nesting. A section inside a section inside a tab gives the user no orientation and pushes the fields into a narrow gutter.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/form-section-pattern
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.