Grabber
Text as a usable component
What happens when we treat text as a usable component?
Text holds a lot of information. Before AI we used text to communicate from human to human, but now text can also direct, instruct, and collaborate with AI models.
This exploration examines the role of text in new workflows and shows how a simple interaction, highlighting and grabbing text, can lead to three different types of outputs:
- Ask AI: drop text into an AI chat → becomes the prompt or context for the model's response.
- Create a task: drop text into a task bucket → becomes a task.
- Breathe: drop text into a mindfulness bucket → becomes an affirmation for a breathing practice.
Ask AI
The words become the instruction: grabbed text lands in the chat as the prompt or its context.
Create a task
The task carries its context: the text arrives as the task itself, never retyped into a to-do app.
Breathe
Text as material for mindfulness: the dropped line comes back as an affirmation to breathe with.
What shaped the work
- Designed and prototyped solo in July 2026, with Claude Code as the only engineering resource.
- Text only by design: the interaction works exclusively with selected text, not images, files, or structured objects.
- One interaction at a time: highlighting surfaces a single action, and everything else waits behind the drop.
- No usage data yet: an unreleased concept, so decisions leaned on interaction principles like progressive disclosure and intentional friction rather than analytics.
One action on highlight
Work solely with text and limit the experience to a single interaction at a time to answer the question, "What can I do with this text right now?"
Rejected: a multi-option toolbar on highlight. It front-loads choice the moment text is selected and turns a simple selection into reading a menu.
Primitive: the grab. Selected text becomes a component you physically pick up, which makes the next step feel like handling material instead of issuing a command.
Buckets as gradual disclosure
By adding light friction, this interaction encourages intentional decision-making. Highlighting text triggers a single, clear action, dropping it into a bucket, rather than a chaotic menu. This progressive disclosure treats text as a flexible component, preventing cognitive overload and driving purposeful engagement.
Rejected: zero-friction instant actions. Removing every step makes each outcome accidental instead of chosen; the slight friction is what makes the action purposeful.
Primitive: the bucket. A named drop target that tells the text what it is about to become, and gives the user a beat to decide.
A wall of options vs one clear next action
The difference in decision making. A toolbar that surfaces a whole set of options the moment text is highlighted, against a single clear next action to take.
Reducing choice is what makes text usable as material.
Treating text as a component only works if the first step is obvious. Limiting highlight to one action removed the cognitive load a toolbar piles onto every selection, and moving the decision to the drop made each outcome feel chosen rather than clicked. The next signal to chase is daily-use frequency: whether the friction still feels purposeful on the fiftieth grab, and whether the buckets explain themselves without a first-run tour.