Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.gumnut.ai/llms.txt

Use this file to discover all available pages before exploring further.

Gumnut’s Immich compatibility layer supports most of the core Immich functionality. This page covers what’s supported, what isn’t, and how Gumnut differs from the official Immich server.

Supported Features

These Immich features work with Gumnut:
FeatureDetails
Photo and video uploadUpload from mobile app, web app, or API
Background backupAutomatic camera roll backup from the mobile app
Photo browsingTimeline view, thumbnails, full-resolution viewing
AlbumsCreate, edit, delete albums; add and remove assets
People and facesFace detection, person tagging, person management
SearchSemantic search (natural language) and metadata search
TimelineMonthly time buckets with date filtering
SyncIncremental sync between devices with checkpoint management
MetadataEXIF data extraction, dates, camera information
Real-time updatesWebSocket notifications for uploads and deletes
AuthenticationLog in with your Gumnut credentials
TrashSoft-deleted assets are recoverable for 90 days before permanent deletion

Not Yet Supported

These Immich features are not currently available through Gumnut. Some may be added in the future:
FeatureStatus
TagsReturns empty results
Maps and placesLocation-based browsing not available
MemoriesAuto-generated memory collections not available
User sharing and partnersSharing libraries between users not available
Shared linksPublic sharing links not available
DuplicatesDuplicate detection view not available
StacksImage stacking not available
NotificationsPush notifications not available
Features listed as “not yet supported” return empty results in the Immich apps — they won’t cause errors, but the corresponding sections will appear empty.

Intentionally Different

Some features work differently in Gumnut by design:
  • No external libraries — Immich allows importing photos from external file paths on the server. Gumnut does not support this because it runs as a managed cloud service, not on your local machine. You always have full access to your original assets and metadata through the API.
  • No admin-level ML configuration — Immich lets you customize embedding models, ML processing settings, and other server-level options. Gumnut manages these automatically — you get the latest models and optimizations without configuration.
  • More accurate face detection and people grouping — Gumnut uses a more sophisticated recognition pipeline than Immich’s per-face matching, with crowd-face filtering, library-wide re-clustering, sibling-aware matching, and protected named Persons. See Faces and People for details.
  • CDN delivery instead of home-server serving — A self-hosted Immich server serves photos directly from your home machine and network, which is limited by your home upload bandwidth and requires exposing the server to the internet. Gumnut delivers your photos through Cloudflare’s global CDN backed by cloud storage, so they load quickly from anywhere without those constraints.
  • No server administration — Immich exposes admin endpoints for managing users, jobs, and system settings. Gumnut handles infrastructure and administration automatically.

Gumnut vs Official Immich Server

FeatureGumnutImmich Server
MCP server + MCP AppYesNo
AI agent integration (Claude, ChatGPT)YesNo
AI description generationYesNo
Photo quality scoringYesNo
Custom embedding modelsNo (managed by Gumnut)Yes
External librariesNo (API access to originals)Yes
Horizontal scalabilityYes (cloud-native)No
CDN for image servingYes (Cloudflare)No
StorageS3-compatible cloudLocal disk
HostingManaged cloudSelf-hosted