Memory is the product. Everything else is an interface.

Most of what we capture in a day never comes back. A voice memo at a red light. A photo of a doctor's prescription. Three lines for a thesis intro. A name we promised to remember.

It all goes into apps that were never designed to bring it back at the right moment. Notes in one place. Reminders in another. Voice memos somewhere on the phone. A folder you can't find anymore. By the time you need any of it, you don't remember where you put it — or that you saved it at all.

It isn't a memory problem. It is a tooling problem. Our tools store. They almost never return.

A few years ago, after a particularly chaotic week, I realized I had been writing the same idea down for the third time in three different apps. The first time on a sticky note. The second on a Telegram chat with myself. The third as a voice memo I never relistened to.

The idea was good. The problem was not the idea. The problem was that my tools could not bring it back to me at the moment it mattered.

That morning, the question changed. It stopped being "what should I capture this in?" and became "why does the act of capture not lead to memory?"

From that question, wamid started.

Notes, reminders, voice memos, photos, half-thoughts — these are interfaces. The product is the memory underneath : structured, sourced, governable, portable. The one you can rely on tomorrow, in six months, in three years.

The model can be swapped. The provider can change. The interface can be redesigned. The memory has to stay yours, in plain text, exportable, durable across decades.

This is why wamid is built around memory as a product — not chat, not summaries, not graph views, not folders. Memory first. Everything else is a way to feed it or get something back from it.

Almost every productivity app converges to the same shape. A chat box. A graph view. A daily digest. A streak counter. Notifications that demand your attention.

We refused all of them. Not because they are bad, but because they were the loudest things we could build — and the loudest things you can build are rarely the most useful.

No chat as the main interface. No infinite scroll. No notification spam. No background recording. No data sold. No training on your captures. No exception buried in a legal paragraph nobody reads.

Every refusal narrowed the design until what remained was the actual product : a memory that comes back.

One person leads the product. A small circle of trusted collaborators and ten early testers shape it day by day. The Android app is built by hand, in Kotlin, on the same phone the founder uses every morning.

The AI runs through our own gateway with Zero Data Retention enforced — no OpenAI in the data path, no Anthropic, no Google. The reasoning model is open. The embeddings are on-device. The sensitive captures never leave the phone, even on the Pro tier.

This is not optimal for growth. It is optimal for trust. Every architectural choice was made by asking : would I let someone I love store their hardest memories in this product ?

wamid (وميض) is the Arabic word for a steady glimmer. The light that doesn't go out. The container is calm, stable, present.

Wisp is the spark that brings back what you forgot. The assistant is sharp, contextual, and arrives only when something deserves to come back.

Return Engine is the silent system underneath. It scores dormancy, decides what matters, and surfaces the right capture at the right moment — without dashboards, without you ever asking.

Three pieces. One promise. You capture. wamid remembers. Wisp brings it back.