DonutMediaCenter · formerly LocalBooru

Your media. Your library.

Browse Images, Videos and Music in dedicated views, connect your private libraries and add the tools you need.

FREE — Open source

Download DonutMediaCenter 2.1.0 View GitHub Explore add-ons

Release and development

DonutMediaCenter 2.1.0 is available.

The first DonutMediaCenter release brings dedicated Images, Videos and Music views, connected-library address discovery and the optional Donut Create add-on. Existing LocalBooru libraries and pairings remain compatible.

Linux and Windows packages are available, alongside a macOS preview and experimental Android packages. Package checks passed; additional runtime acceptance was waived for this release. Creator commerce and wider public discovery remain future work.

Read the 2.1.0 release notes and platform limitations.

The library

Keep the core focused

The ordinary library runs separately from optional processing add-ons. Install the extra tools only when you need them.

Local folders

Index and browse the media you choose to add. Optional processing has its own dependencies.

Media-specific browsing

Images, Videos and Music are the development direction, with relevant filters and gallery-scoped collections.

Explicit sharing

Private device connections and creator publication are separate relationships. Connecting a device should not imply permission to publish its library.

Choose your tools

Add the tools your library needs

Choose optional processing tools alongside the core library. Each add-on has its own setup requirements. Donut Create is currently in development.

2.0.5 add-on · optional install

Auto Tagger

Create searchable image tags with local ONNX classification models. Review the suggestions and choose the thresholds that fit your collection.

  • Separate processing add-on
  • Model files and Python environment required
  • Tags are model estimates, not authoritative descriptions
2.0.5 add-on · optional install

Whisper Subtitles

Generate subtitle text from the audio in your videos, then check names, timing and any difficult passages.

  • Whisper-based speech recognition
  • Model and processing dependencies required
  • Review transcription before sharing it
2.0.5 add-on · optional install

Chromecast/DLNA

Send supported media to compatible players on your local network. Device discovery and playback support depend on the receiver and media format.

  • Chromecast and DLNA integration
  • Compatible receiver and reachable local network required
  • Available formats depend on your device
2.0.5 add-on · external setup required

SVP frame interpolation

Use a compatible SVP and VapourSynth setup for an optional interpolated-video playback path. Platform support and responsiveness need checking on the intended machine.

  • Requires compatible external SVP/VapourSynth components
  • Hardware and video format affect performance
  • Keep ordinary playback available when the add-on cannot run
2.0.5 add-on · managed bundle

LADA Video Restoration

An optional accelerated video-restoration add-on with its own managed runtime and model requirements. Check source suitability and the processing result before relying on it.

  • Installed separately from the core library
  • Compatible acceleration/runtime and models required
  • Use only media you have the right to process
2.0.5 add-on · optional install

Age Detector

Generate model-based image and face estimates for library filtering. Estimates can be wrong and do not verify anyone's age or replace manual review.

  • Separate local machine-learning add-on
  • Additional models and dependencies required
  • Not age verification or a guarantee of content safety
2.0.5 · built-in

Curation Game

Work through the current gallery to make keep-or-discard decisions without changing the scope of the library you are reviewing.

  • Managed as a built-in add-on
  • No separate Python process
  • Review removal behavior before acting on irreplaceable files
2.0.5 · built-in

WD14 Text Sidecars

Import and export same-stem tag text files so image annotations can move between tools without being locked into one database.

  • Import, absorb and export tag sidecars
  • Managed as a built-in add-on
  • No separate Python process
In development

Donut Create

A creation workflow connecting DonutUI, ComfyUI and DonutNodes. It is being integrated into DonutMediaCenter and should not be assumed to ship in the current 2.0.5 downloads.

  • Current development work
  • Separate ComfyUI runtime and model setup
  • Release availability and end-to-end acceptance still need confirmation

Download DonutMediaCenter 2.1.0

Linux x86-64 requires glibc 2.39 or newer. Windows packages are unsigned. macOS is a non-notarized preview; Android cold-start acceptance remains incomplete. See the release notes for details.