← All projects
Off Featured

Quid

Notes and tasks as the same object

A workspace where notes and tasks are the same thing in two states. Activate a note and AI breaks it down into concrete, actionable tasks.

Role
Full Stack Developer & Interaction Designer
Year
2026
Status
Off
Read
2 min
Stack
Next.jsConvexGemini 1.5 FlashFramer Motion
The story

The problem

Most people keep a 'note graveyard' and a separate task list, with almost nothing connecting the two. Ideas stay ideas and never turn into action.

The approach

One engine for both: an item is a note until you activate it. On activation, Gemini reads the note and proposes concrete tasks.

My role

Product concept, interaction design and full-stack build.

Where it stands

No longer online. The MVP ran with a public demo and was tested on my own workflow.

Next step

Add a 'gatekeeper' mode that helps decide which notes deserve to become tasks, to prevent task inflation.

The project

The idea

Quid started from a simple observation: notes and tasks live in different places, and very little connects them. Notes are potential energy: ideas, snippets, inspirations. Tasks are kinetic energy: actions, deadlines, commitments.

I wanted an app with no distinction in the engine, only in the state. An item is a note until you activate it, and activation is the moment an idea becomes work.

What it does

  • Capture freely: everything starts as a note, with no structure required.
  • Activate: when a note is ready, one action turns it into tasks.
  • AI breakdown: Gemini reads a messy note, for example about “fixing the garden”, and proposes concrete steps: “Buy mulch”, “Prune the roses”, “Check irrigation timing”.
  • Feel the shift: the transition from note to task is animated to feel weighty and permanent, not like ticking a box.

How I built it

The build focused on speed and materiality:

  • Stack: Next.js 14 and Convex for a fully reactive, real-time interface.
  • AI as a parser, not a chatbot: Gemini 1.5 Flash is used as a structural parser that turns free text into a list of actionable tasks.
  • Design: high-contrast and dark-first, with a single purple accent (#8B5CF6) that marks the energy shift during activation.
  • Micro-interactions: Framer Motion animates the move of a note into the task “level”.

What I learned

The AI proposals were most useful when I was tired, exactly when breaking things down is hardest. But the real lesson was another one: activation is a cognitive tax. Making it easy to turn a note into a task doesn’t mean I actually want to do that task.

I started from the idea that more tasks means more productivity. In practice Quid made it too easy to fill my list with low-value actions born from incidental notes.

Treating notes and tasks as one object is technically sound and feels elegant. But the interface has to guard against task inflation: AI is great at breaking things down, and a human still needs to decide what is worth doing.

Other experiments
Got an idea?

Let's talk.

Collaborations, product challenges, or just a few thoughts on AI and delivery.

Contact me →