Roadmap

What's next, and what never will be

No dates. A missed public date costs more than a vague one, and this is one person's work alongside other work. Ordering changes based on what people actually ask for.

Shipped

Zero-config folder watching and sub-second ingest
On-device Apple Vision recognition, with screen-text and handwriting profiles
Global overlay search with typo-tolerant and meaning fallbacks
Library with filters, tags and live smart collections
Redaction with automatic secret detection
Resumable backlog import
Retention rules, recoverable Trash, near-duplicate review
Optional AES-GCM encryption of the recognised-text index
Send to Reminders, Things, Obsidian, Apple Notes and the share sheet

Full detail in the changelog.

Shortcuts and AppleScript actionsExpose search, copy and export as Shortcuts actions so the library can be driven from your own automations. The service layer already exists internally.

Move the library to another diskCurrently fixed under Application Support. People with large libraries and small internal drives want it elsewhere, and that's fair.

More recognition languages, properly testedThe storage and index are already language-agnostic. What's missing is real testing beyond English before claiming support.

Considering

Genuinely undecided. If one of these is the thing standing between you and buying, saying so moves it.

iCloud sync between your own MacsThe hard part isn't sync, it's conflict resolution without a server to arbitrate. Stable capture identifiers exist from day one so it stays possible. Would never mean uploading images to us — there is no us to upload to.

An alternate engine for handwritingVision is weaker on handwriting than on screen text. Only worth it if enough people are actually capturing whiteboards, and only with an engine whose licence suits a paid closed-source app.

SetappA second channel reaching people who'd never find this site. The trade is revenue share and losing the direct relationship with customers.

Not doing

These are decisions rather than a queue. A small app stays good by refusing things, and it's more useful to you to know now than to wait for a feature that isn't coming.

Cloud sync or team librariesTurns a local tool into a backend business with authentication, billing tiers and uptime obligations — and directly contradicts the one promise the app makes.

Cloud AI summarisation of screenshotsWould require uploading your images. Same reason.

Cloud OCR fallback when recognition is poorThe failure case would be exactly the sensitive document you most wanted kept local. A local miss stays local.

An annotation and markup editorCleanShot X and Preview do this well. Competing there is a feature war against funded teams, fought by making this app bigger and worse.

Screen recording, GIF or video captureA different product with a different storage problem and a much larger surface area.

An iOS or iPad companionDoubles the work, and the Desktop-clutter problem is Mac-only.

Analytics, telemetry or crash reportingNot even opt-in, not in v1. It would make the central claim negotiable, and the moment it's negotiable it isn't a claim.

Arguing with this list