Best Self-hosted Alternatives to Amazon Photos

Immich is an open-source photo and video management platform that provides web and mobile clients, background processing, and searchable metadata. It focuses on efficient ingestion, indexing, and playback of large personal or small-team media libraries while supporting S3-compatible storage backends.

Key Features

• Web and mobile clients (web UI + native Flutter mobile apps) for browsing, backup, and playback
• Background job pipeline for thumbnails, transcoding, metadata extraction, and duplicate detection
• Face detection, face clustering, and visual/object search capabilities (embeddings-based search)
• Search by metadata, EXIF/location, objects, and faces; global map browsing of geotagged assets
• Support for S3-compatible storage backends and local filesystem storage abstractions
• Multi-user accounts, sharing, albums, and access controls with API and SDK support
• Queueing and worker architecture for scalable processing (Redis-backed queues) and database migrations
• Container-first deployment with Docker and orchestration-friendly configuration

Use Cases

• Automated mobile backup and centralized management of personal photo/video collections across devices
• Small teams or families sharing and organizing media with searchable tags, faces, and map-based browsing
• Photographers and hobbyists who need a private catalog with metadata search, RAW/HEIC support, and transcoding

Limitations and Considerations

• Machine-learning features (face detection, embedding generation, CLIP/object search) can be CPU- and memory-intensive; large libraries may benefit from dedicated workers or GPU resources
• Some features rely on database capabilities and indexes that may require specific PostgreSQL versions or extensions and careful migration planning
• Native dependencies used for image/video processing and transpilation (e.g., FFmpeg and native image libraries) can increase build/installation complexity on some platforms

Immich provides a full-featured alternative for private media management with strong indexing and mobile-first backup. It is maintained as an open-source project with an active community and a focus on performant ingestion, search, and sharing for photo and video libraries.

Jellyfin is a free software media system for collecting, organizing, and streaming your personal media library from your own server to many types of clients. It provides a server backend and API along with a web interface, and is commonly used as an open alternative to proprietary media servers.

Key Features

• Library management for movies, TV shows, music, and photos with metadata fetching
• Web-based administration and playback interface, plus a broad ecosystem of official and third-party clients
• Streaming with on-the-fly transcoding support via FFmpeg
• User accounts and profiles for separating access and playback history
• Extensible architecture with plugins and integrations

Use Cases

• Host a private “Netflix-like” server for a household’s movie and TV collection
• Centralize and stream a music library to phones, desktops, and smart TVs
• Provide media access for friends or remote devices while keeping content on your own server

Limitations and Considerations

• Transcoding and high-bitrate streaming can require significant CPU/GPU resources depending on usage
• Some client capabilities and codecs may vary by platform, affecting direct play vs transcoding

Jellyfin focuses on giving you full control over your media, with no tracking or vendor-operated central services. It is well-suited for home labs and organizations that want a flexible, privacy-respecting media streaming stack.

PhotoPrism is a self-hosted photo management application that helps you organize, browse, and search large photo and video libraries. It uses machine learning to automatically classify content and make your media easier to find while keeping control of your data.

Key Features

• Automatic labeling and categorization using machine learning
• Face recognition to group photos of people
• Powerful search with combinable filters (labels, places, colors, quality, and more)
• Map-based browsing and location features with geocoding support
• Metadata extraction and merging (e.g., Exif and XMP)
• Web interface optimized for desktop and mobile (PWA-style experience)
• WebDAV support for managing files from compatible clients

Use Cases

• Replace or complement cloud photo services with a private photo library
• Organize multi-device family photo collections with search, labels, and people
• Manage and curate photo/video archives on a home server or NAS

PhotoPrism combines a modern web UI with AI-assisted organization to make personal media libraries searchable and enjoyable to browse. It is well-suited for users who want fast search, automatic tagging, and flexible access to their photo collection.

Ente is an open-source, end-to-end encrypted cloud platform designed to store and sync personal data without needing to trust the service operator. It powers apps such as Ente Photos (a private alternative to mainstream photo clouds) and Ente Auth (a 2FA authenticator with encrypted sync).

Key Features

• End-to-end encryption with a zero-knowledge design for stored and shared content
• Cross-platform clients (mobile, desktop, and web) for photos and authenticator workflows
• Photo backup with background uploads and multi-device sync
• Private sharing, including collaborative albums and link-based sharing with encrypted uploads
• Migration and export workflows to move photo libraries in and out
• External security and cryptography audits reported by the project

Use Cases

• Private photo and video backup as an alternative to Google Photos or iCloud Photos
• Secure family or team photo sharing with collaborative albums
• Encrypted syncing of 2FA tokens across devices (Authy alternative)

Limitations and Considerations

• Some advanced features (for example, semantic search/face features) may depend on the app’s capabilities and server configuration
• Ente Photos is offered as a paid service by the vendor, though the codebase can be deployed independently

Ente suits users who want a privacy-first, audited, end-to-end encrypted experience for photos and 2FA, while keeping the option to run the full stack under their own control. It combines modern client apps with an encrypted cloud backend focused on minimizing trust.

LibrePhotos is a self-hosted photo and video management service that scans your library from disk and helps you browse, search, and organize your media. It includes machine-learning features for richer discovery, such as face recognition and semantic search.

Key Features

• Imports and indexes existing folders by scanning the file system
• Supports photos (including many RAW formats) and videos
• Timeline view for browsing media chronologically
• Multi-user support
• Metadata-based search (EXIF and other embedded information)
• Face detection and face clustering/classification
• Object and scene detection plus image captioning to improve discovery
• Reverse geocoding to infer locations for photos
• Semantic image search

Use Cases

• Replace proprietary photo cloud services with a self-managed library
• Organize large family photo archives with faces, places, and smart search
• Run a multi-user home or small-team media library with individual accounts

Limitations and Considerations

• Machine-learning features can require significant CPU/RAM and indexing time on large libraries
• Advanced features like video/image conversion depend on external tools being available and correctly configured

LibrePhotos is a strong option for anyone who wants a modern, searchable photo library under their own control. It combines classic photo management features with practical AI-assisted organization for easier browsing and discovery.

Photoview is a self-hosted photo and video gallery designed for personal servers and photographer-friendly workflows. It scans media from directories on your server, maps folders to albums, and builds thumbnails for fast browsing while keeping files under your control.

Key Features

• Filesystem-based library scanning with folders mapped to albums
• Automatic thumbnail generation and lazy-loading for performant browsing
• Multi-user accounts with per-user library paths and access controls
• Share albums or individual media via public links, optionally password-protected
• RAW photo support (using Darktable for conversion)
• EXIF metadata extraction and display, including map view when GPS data is present
• Video support with web-optimized transcoding (using FFmpeg)
• Face detection and grouping of photos by person

Use Cases

• Personal or family photo library hosted on a home server or NAS
• Photographer archive browsing with RAW and EXIF-focused workflows
• Private sharing of selected albums with friends/clients via protected links

Limitations and Considerations

• The official iOS app is no longer maintained or available in the App Store
• Requires an initial scan and thumbnail generation that can be time-consuming on very large libraries

Photoview is a good fit if you want a fast, web-based gallery that mirrors your existing folder organization. It focuses on performance, privacy, and photographer-oriented features like RAW, EXIF, and mapping.

Lychee is a free, open-source photo management system that runs on your own server. It lets you upload, organize, browse, and share photos and albums from a fast, app-like web interface.

Key Features

• Photo and album management (create albums, move, rename, describe, delete)
• Fast browsing and full-screen viewing with keyboard navigation
• Search, tags, and “star/important” marking for organization
• Share photos/albums publicly and optionally protect albums with passwords
• Metadata support including EXIF and IPTC display
• Multiple import options (local upload, server-side import, and URL-based import)

Use Cases

• Personal or family photo library with full control over storage and access
• Photographer portfolio delivery via shareable albums
• Small team photo repository for events and documentation

Lychee is a solid choice for users who want an attractive, straightforward photo gallery and management workflow without relying on third-party cloud photo services. It combines practical organization tools with simple sharing controls in a clean web UI.

Memories is a photo and video management application implemented as a Nextcloud app. It provides timeline-first browsing, AI-assisted tagging integration, albums and sharing, map views, bulk metadata editing, and on-demand video transcoding for web-friendly playback.

Key Features

• Timeline view that sorts media by date taken (Exif) and supports fast "rewind" navigation through large libraries.
• AI-assisted automatic tagging via integration with Recognize and Face Recognition Nextcloud apps to group by people and objects.
• Albums and external sharing, including collaborative album editing for multiple Nextcloud users.
• Bulk metadata editing (title, description, GPS, date/time, tags) with most metadata preserved in file Exif headers.
• On-demand video transcoding and HLS packaging for adaptive streaming; supports hardware acceleration (VA-API, NVENC) where available.
• Map view with automatic GPS extraction and reverse geocoding to locate photos on a map.
• Migration helpers for importing from Nextcloud Photos and Google Takeout; optimized for large libraries (tested at high photo counts).

Use Cases

• Personal self-hosted photo library on a Nextcloud instance with privacy-first storage and searchable timelines.
• Small teams or families that want shared, collaborative albums and controlled external sharing without third-party hosting.
• Migrating and organizing large collections (including video) while retaining Exif metadata and enabling web playback via HLS.

Limitations and Considerations

• Requires a Nextcloud instance and recent platform stack (Nextcloud 26+, PHP 8.0+); several features require external system binaries and services (ffmpeg/ffprobe, Imagick) which must be installed and configured.
• AI tagging is provided via external Nextcloud apps and models; enabling full face/object recognition may require separate setup and resources.
• Video transcoding can be resource intensive; hardware acceleration is recommended for performance on large libraries.
• Android client is described as early access; mobile upload relies primarily on Nextcloud mobile apps.

In summary, Memories is a focused, performance-oriented photo gallery app for Nextcloud that emphasizes privacy, metadata preservation, timeline navigation, AI-assisted tagging, and web-friendly video playback. It is designed for users and organizations who want a feature-rich, self-hosted alternative for managing large photo and video collections.

Memories is a fast, modern photo and video management application that runs as a Nextcloud app. It provides a responsive web interface to browse large libraries, organize albums, edit metadata, and stream videos efficiently while keeping files in your existing Nextcloud storage.

Key Features

• Timeline-based browsing using extracted EXIF date/time metadata
• Albums with collaborative organization and sharing (including external sharing)
• Bulk metadata editing (titles, descriptions, dates, GPS, tags)
• Map view using GPS EXIF data, including reverse geocoding for locations
• Video playback with on-demand transcoding and adaptive streaming (HLS)
• Archive workflow to hide selected items from the main timeline
• Integrations for automatic tagging and face/object grouping via Nextcloud AI-related apps
• Performance-oriented indexing designed for very large libraries

Use Cases

• Replace or enhance Nextcloud Photos with a faster, feature-rich gallery experience
• Organize and share family or team media libraries through Nextcloud accounts and public links
• Browse and stream mixed photo/video collections on mobile and desktop with efficient transcoding

Limitations and Considerations

• Some advanced features (for example AI tagging) depend on installing and configuring compatible Nextcloud apps
• Video transcoding can be resource-intensive and may require additional configuration for hardware acceleration

Memories is well-suited for users who already rely on Nextcloud and want a dedicated, high-performance media library experience. It emphasizes retaining your existing folder structure and storing key metadata in standard formats to reduce lock-in and ease migration.

PiGallery2 is a fast, directory-first web photo gallery that renders your existing folder structure as a read-only gallery. It is optimized for low-resource servers and focuses on quick browsing, searching, and viewing photos and videos without modifying your originals.

Key Features

• Directory-first browsing that mirrors your on-disk folder structure
• Fast indexing to a database for responsive browsing (SQLite or MySQL)
• Advanced search with boolean logic, negation, range queries, and autocomplete suggestions
• Photo viewer with optional info panel (including EXIF metadata)
• Map view from GPS photo metadata, with GPX track display and activity visualization
• Sharing links for folders with optional password protection
• On-the-fly thumbnail generation in multiple sizes with caching to a temp folder
• Video playback support and optional transcoding with thumbnail generation
• Markdown files in folders for simple blogging/notes and date attachment
• Logical albums based on saved searches and per-user allow/block filters

Use Cases

• Self-host a lightweight family photo archive that keeps your existing folder layout
• Quickly search large photo collections by date, rating, people metadata, and location
• Share selected folders externally via password-protected links

Limitations and Considerations

• The gallery is intentionally read-only and does not support organizing, editing, or tagging photos within the app
• Face features rely on existing metadata (no built-in ML-based automatic face detection)

PiGallery2 is best suited for users who want a fast, minimal-maintenance photo gallery that reflects their filesystem as-is. It provides powerful search and viewing features while keeping photo management in your regular file workflow.

Damselfly is a server-based digital photo management system built to index very large, folder-based libraries and make images easy to find using filenames, folders, and embedded metadata. It emphasizes fast search and an efficient keyword-tagging workflow, with optional computer-vision assistance.

Key Features

• Indexes large photo collections using EXIF/IPTC metadata plus folder and filename information
• Fast full-text search, including multi-phrase and partial-word matching
• AI-assisted tagging with face detection, face recognition, and object detection (runs locally/offline)
• Advanced filters (date ranges, camera/lens metadata, orientation, file size, untagged images, and more)
• Non-destructive keyword tagging via ExifTool to avoid JPEG re-encoding
• Background scanning and automated thumbnail generation
• Multi-user accounts with role-based entitlements (including read-only access)
• Basket workflow to collect images for download/export and further processing
• Optional desktop client for syncing selected images to a local folder for editing

Use Cases

• Personal or family photo library management with quick search and tagging
• Photographer workflows for cataloging, keywording, and exporting selected sets
• Home server/NAS-based photo access across multiple devices without copying catalogs

Limitations and Considerations

• Primarily designed for folder-based photo libraries; edit/manipulation features are limited compared to dedicated editors
• Video support is not a primary focus and may be limited depending on format and workflow

Damselfly is well-suited for users who want a self-hosted, high-performance photo catalog focused on search, metadata, and assisted tagging. Its basket-and-sync workflow also supports bringing selected photos to a workstation for external editing while keeping the main library centralized.

ChronoFrame is a self-hosted personal photo gallery for uploading, organizing, and viewing photos in a modern, responsive web interface. It focuses on smooth browsing (including large images) and metadata-aware organization with map-based exploration.

Key Features

• Web-based photo upload and management with albums and a dashboard
• EXIF parsing to extract capture time, GPS coordinates, and camera details
• Reverse geocoding to identify shooting locations and map-based exploration view
• Live Photo and Motion Photo support (pairing image and MOV video components)
• Multiple storage backends, including local filesystem and S3-compatible storage
• SQLite-based setup option for low-maintenance deployments

Use Cases

• Personal or family photo library with location-aware browsing
• Lightweight self-hosted alternative for showcasing a frequently updated gallery
• Curating trips or events with albums and map exploration

ChronoFrame is a strong fit for users who want an easy-to-deploy, metadata-driven gallery with modern UI performance and support for dynamic photo formats. Its combination of map exploration, EXIF tooling, and flexible storage makes it practical for both private archiving and sharing curated collections.

Slink is a self-hosted, open-source image sharing platform that lets teams and individuals upload, manage, and share images from their own instance. It pairs a Symfony backend with a SvelteKit frontend and is distributed with Docker images and a demo instance for evaluation.

Key Features

• Multi-format image uploads with automatic processing (PNG, JPG, WEBP, SVG, BMP, ICO, GIF, AVIF, HEIC, TIFF) and format conversions where required.
• Multi-file upload UI with per-file progress and upload history (list and grid views).
• Hierarchical (nested) tags and search/filtering for organized image discovery.
• Image deduplication to prevent duplicate storage and user-facing notifications on duplicates.
• Shareable/customizable image links, URL shortening for images, and ShareX integration for automated uploads.
• Role-based authentication, optional user approval workflow, admin dashboard with analytics, notifications, bookmarking, and public gallery browsing.
• Docker-ready deployment with docker run / docker-compose examples; supports local storage, SMB shares, and S3-compatible backends.

Use Cases

• Host and share project screenshots, portfolio images, or documentation assets without relying on third-party image hosts.
• Provide a team-facing image upload endpoint (ShareX/API key integration) for automated capture and sharing workflows.
• Public or community galleries for artists or small communities with moderation and bookmarking features.

Limitations and Considerations

• Public API exists but is not fully documented; integrations may require reading source code or opening issues for guidance.
• HEIC and TIFF images may be converted to JPG for compatibility, which can change metadata and image fidelity.
• Advanced production concerns (CDN integration, multi-node database scaling) are listed as upcoming or optional features; plan infrastructure accordingly for large-scale usage.

Slink provides a pragmatic, developer-friendly way to self-host image sharing with a modern Symfony + SvelteKit stack and container-first deployment options. It is suitable for single-server homelabs, small teams, and anyone who wants control over image hosting and sharing policies.

Immich Kiosk is a lightweight web application for creating highly configurable slideshows from an Immich photo library. It’s designed to run in browsers and kiosk-style displays (such as tablets, smart displays, or a small PC connected to a TV) using Immich as the media source.

Key Features

• Displays Immich assets as full-screen slideshows optimized for kiosk devices and browsers
• Highly configurable behavior and presentation (slideshow style, selection rules, and display options)
• Configuration via URL query parameters for easy remote control and simple device setups
• Responsive, device-friendly UI suitable for different screen sizes and orientations
• Container-friendly deployment with straightforward Docker-based installation

Use Cases

• Digital photo frame for households using Immich as their primary photo library
• Wall display/TV slideshow for events, offices, or shared spaces powered by an Immich server
• Dedicated kiosk mode display on tablets or low-power devices showing curated Immich albums

Limitations and Considerations

• Not affiliated with the Immich project and relies on Immich availability and API compatibility

Immich Kiosk is a practical add-on for Immich users who want a polished, configurable slideshow experience on always-on displays. It focuses on simple deployment and flexible configuration while leveraging Immich as the single source of truth for media.

bewCloud is an open-source personal cloud platform designed as a simpler, lighter alternative to Nextcloud and ownCloud. It provides a modern web UI for keeping your personal files and everyday cloud features under your control, with a focus on low resource usage and minimal complexity.

Key Features

• Web-based file storage and browsing
• Built-in apps for photos, notes, and news/feeds
• Optional CalDAV/CardDAV support for calendars and contacts (via Radicale integration)
• Database migrations and background cron jobs for maintenance tasks
• Containerized deployment support for straightforward hosting

Use Cases

• Replace a heavier all-in-one cloud suite for a personal or family cloud
• Centralize files, notes, photos, and feeds on a private server
• Run a lightweight home server cloud with optional calendar/contact sync

Limitations and Considerations

• Admin hardening is required: the first signup can become an admin even if signups are disabled
• CalDAV/CardDAV functionality depends on additional Radicale configuration

bewCloud is a strong fit if you want a modern, streamlined web cloud with fewer moving parts than traditional suites. It prioritizes simplicity and efficiency while still covering common personal cloud needs.

HomeGallery is a self-hosted, open-source web gallery for browsing and exploring personal photo and video archives. It focuses on fast, mobile-friendly navigation, tagging, and AI-assisted discovery such as similar-image search and face search.

Key Features

• Timeline view and similarity-based browsing for rediscovering related photos
• Reverse image lookup (similar image search) to find visually related content
• Face detection and finding similar faces
• Video preview generation and transcoding
• Tagging with single and multi-selection
• Boolean search/query language with and/or/not operators
• Reverse geocoding (geo coordinates to human-readable locations)
• Static site export to run as a standalone gallery without a backend service
• Supports offline/read-only media sources after previews and metadata are generated

Use Cases

• Create a private, searchable family photo and video archive on a NAS or home server
• Explore large media collections using similarity browsing to surface forgotten memories
• Generate a static web gallery export for sharing a curated collection internally

Limitations and Considerations

• The full media index is loaded into the browser, which can become large for very big libraries
• Some AI capabilities may rely on an API server that can be self-hosted if you want to avoid external calls

HomeGallery is well-suited for individuals who want a privacy-preserving alternative to cloud photo services while keeping a responsive experience on mobile devices. Its similarity-based exploration and client-side searching make it especially useful for large, long-lived media archives.

WeddingShare is a lightweight self-hosted web application for collecting and sharing photos before, during, and after a wedding. It lets you create one or more galleries with shareable links so guests can easily upload and view memories.

Key Features

• Create multiple galleries, each with its own shareable link
• Guest-friendly interface for viewing and uploading photos
• QR code generation for easy sharing on invites or table cards
• Optional review/moderation flow to approve uploads before they appear
• Multiple viewing modes, including presentation mode and slideshow mode
• Configurable allowed upload file types via environment settings

Use Cases

• Collect guest photos from a wedding day into a single shared gallery
• Run a live slideshow or projected gallery during the reception
• Share pre-wedding planning moments (venue visits, fittings, tastings) with friends and family

Limitations and Considerations

• Galleries are link-accessible; anyone with the link can view and upload, and the optional secret key is only a basic deterrent
• Some image formats may not be supported by browsers even if uploads are allowed (for example, HEIC)

WeddingShare is a practical option for couples who want a simple, private alternative to social platforms for wedding photo sharing. Its QR-based sharing and display modes make it especially suitable for in-person events.

Meme Search is an open source meme search engine and finder that helps you organize and retrieve images using AI-generated descriptions and semantic vector search. It performs image-to-text extraction, embedding generation, and searching locally, designed to run easily via Docker.

Key Features

• Local image-to-text processing with multiple selectable captioning models
• Automatic and manual meme description management to improve search quality
• Fast semantic and keyword search backed by PostgreSQL with pgvector
• Tagging and filtering by tags, directory paths, and search mode (keyword vs vector)
• Bulk description generation for faster indexing
• Directory rescans to detect and index newly added images
• Drag-and-drop uploads via the web UI (with configurable storage directory)
• Dark mode user interface

Use Cases

• Personal meme library management for quick semantic retrieval by content or text
• Homelab media organization with local-first AI indexing and search
• Team/shared collections where tags and edited descriptions improve discoverability

Limitations and Considerations

• First-time caption generation can be slow due to downloading and caching model weights
• Resource usage depends on the chosen image-to-text model size and hardware capabilities

Meme Search combines a Rails-based web interface with a Python-based generator to provide practical, private semantic search for image collections. It is well-suited for users who want fast local search, tagging, and ongoing rescans without relying on external AI services.

Photofield is a self-hosted photo gallery and viewer designed for very fast browsing of large photo libraries on commodity hardware. It is non-invasive: your filesystem is the source of truth, and originals are not modified.

Key Features

• Seamless zoomable interface across views for quick detail inspection
• Progressive multi-resolution loading from previews to full-quality media
• Read-only, filesystem-based collections with non-destructive behavior
• Fast indexing optimized for large libraries
• Flexible thumbnail and media pipeline, including reuse of existing thumbnails
• Basic video support (including use of pre-generated multiple resolutions)
• Optional semantic search via a separate AI component and optional tagging (alpha)

Use Cases

• Browsing and searching large personal photo archives stored on NAS or local disks
• Complementing other photo management tools with a fast, read-only viewer
• Serving a lightweight family photo portal with directory-based separation

Limitations and Considerations

• No built-in user accounts, authentication, or authorization
• Not optimized for many concurrent users due to server-side state handling
• Initial page load can be slow on weaker CPUs or cold storage caches
• No on-the-fly video transcoding

Photofield is a strong fit for users who want a fast, simple, non-destructive photo viewer that scales well to large libraries. Its single-binary deployment and caching approach make it easy to run while keeping originals untouched.

Omoide is a self-hosted, offline-first photo and video library that organizes media locally using on-device AI models. It focuses on privacy and longevity by keeping all processing and search on the user’s machine, and can be run as a desktop app or in Docker for servers and NAS devices.

Key Features

• 100% local / offline-first media management (no cloud required).
• Face detection, embedding and clustering with review/merge tools for people management.
• Semantic (natural-language) image search powered by OpenCLIP embeddings.
• Auto-tagging, perceptual duplicate detection, and video scene extraction for faster curation.
• Map view with EXIF extraction and GPS editing; co-appearance graph to visualize relationships between people.
• Multiple profiles, read-only mode for sharing, and background task management for large libraries.

Use Cases

• Personal archival and private photo/video libraries where users want local AI-powered search and face grouping.
• Photographers or hobbyists organizing large collections on a NAS or local server (Docker support).
• Small galleries or archivists needing an offline, searchable presentation/preview of media with privacy controls.

Limitations and Considerations

• License: distributed under the PolyForm Noncommercial License 1.0.0, which restricts commercial use; review the license before deploying in commercial contexts.
• Resource & setup: initial model downloads and AI processing can require significant CPU/GPU resources; FFmpeg and Python toolchain are required for desktop builds and media processing. Performance and required hardware depend on chosen models (CPU-only runs will be slower).
• Community & support: actively maintained as a passion project; community discussion and feedback appear on self-hosting forums and social channels rather than a dedicated commercial support portal.

Omoide provides a privacy-focused, local-first alternative for people who want Google/Apple Photos–like discovery features without cloud dependencies. It is suited for users who can accommodate model downloads and local compute demands and who accept the project’s noncommercial license terms.