Snapwell — Your inspiration, organized by AI

Collect anything, search by what you see. An open-source inspiration library for macOS and iOS.

Built by @gustavscirulis.

How it works

1. Import images and videos — Screenshots, photos, video, web images. Drag in, paste, or share from any app.
2. AI tags everything — Every image gets analyzed automatically. Add guidance to focus on what matters to you.
3. Search by what you see — Describe an object, a layout, a mood. Snapwell finds the match — no tags or folders needed.

Built for visual thinkers

• UI design — UI patterns, login flows, component states, animations. Build a searchable reference library of interface design.
• Home improvements — Interior design ideas, furniture, paint swatches, room layouts. Organize into spaces per project.
• Creative inspiration — Art, typography, photography, fashion, video references. Let AI find connections across your collection.

Features

• On-device storage — All images, metadata, and preferences stay on your devices and your iCloud — not our servers
• AI-powered analysis — Categorize and tag using OpenAI, Claude, Google Gemini, or OpenRouter with your own API key
• Custom analysis prompts — Configure AI instructions per space for tailored categorization
• Spaces — Organize images into collections with drag-and-drop and per-space export
• Search by content — Find images based on what AI detected in them, not filenames
• iCloud sync — Sync your library between Mac and iOS
• iOS Share Extension — Send images and videos to Snapwell from any app
• Video support — Import and analyze video alongside images

Privacy

No subscription. No account. Open source.

Your library lives on your devices and your iCloud — not our servers. Snapwell has zero analytics, zero telemetry, and zero tracking. AI analysis goes directly from your device to the provider you choose. You supply your own API key. See PRIVACY.md for details.

Installation

Download from the App Store (Mac and iOS), or build from source.

Requirements

AI analysis requires an API key for at least one provider: OpenAI, Anthropic (Claude), Google Gemini, or OpenRouter. Add your key in Settings and pick your preferred model. The app works without AI too — you just won't get automatic tagging and search.

File storage

Snapwell stores files in ~/Documents/Snapwell/ or iCloud Drive:

• images/ — Media files (PNG, MP4, etc.)
• metadata/ — JSON sidecar for each media item
• thumbnails/ — Generated thumbnails
• spaces.json — Space definitions and AI prompt configuration
• .trash/ — Deleted items (auto-emptied after 30 days)
• queue/ — iOS import staging, auto-watched by the Mac app

Development

Snapwell is built with Swift 6, SwiftUI, and SwiftData. The Mac app uses XcodeGen for project generation.

macOS

cd macos
xcodegen generate
open Snapwell.xcodeproj

iOS

open ios/Snapwell.xcodeproj

Running tests

# Mac
cd macos && xcodegen generate && xcodebuild test \
  -project Snapwell.xcodeproj -scheme Snapwell \
  -destination 'platform=macOS' 2>&1 | xcbeautify --quiet

# iOS
cd ios && xcodebuild test \
  -project Snapwell.xcodeproj -scheme Snapwell \
  -destination 'platform=iOS Simulator,name=iPhone 17 Pro' 2>&1 | xcbeautify --quiet

Contributing

Contributions welcome — open a Pull Request or file an issue.

License

This project is licensed under the GNU General Public License v3.0 — see LICENSE for details. As an additional permission under GPL-3.0 section 7, this software may be distributed through the Apple App Store.

Acknowledgments

• Built entirely with Claude Code by Anthropic
• App icon by Midjourney
• AI analysis powered by OpenAI, Anthropic, Google Gemini, and OpenRouter