Releases: neurallab-dev/mediafind-releases
MediaFind 0.1.7
Native desktop installers for Windows and Ubuntu/Linux (unsigned beta), plus standalone server binaries.
Recommended — native installers (double-click)
- Windows:
MediaFind_0.1.7_x64-setup.exe(NSIS installer) - Ubuntu / Linux:
MediaFind_0.1.7_amd64.AppImage(portable) orMediaFind_0.1.7_amd64.deb
The OS may warn on first launch (these installers are not code-signed).
Advanced — standalone server binaries (terminal)
MediaFind-0.1.7-linux-x64.tar.gz/MediaFind-0.1.7-windows-x64.zip— extract and runmediafind serve.
macOS ships separately as a signed & notarized .dmg.
MediaFind 0.1.5
Unsigned beta server binaries for Windows (x64) and Ubuntu/Linux (x64).
These are PyInstaller builds of the mediafind server, run from a terminal. They are not code-signed — your OS may warn on first launch.
- Windows: extract
MediaFind-0.1.5-windows-x64.zip, then runmediafind.exe serve - Linux: extract
MediaFind-0.1.5-linux-x64.tar.gz, then./mediafind serve
Then open the printed http://127.0.0.1:<port> URL.
The macOS app remains the signed & notarized .dmg (see the 0.1.4 release).
MediaFind 0.1.4
MediaFind 0.1.4 — local, private transcription + natural-language search over your audio & video. macOS (Apple Silicon), signed & notarized.
Highlights
- Redesigned Home / library browser — a dedicated Home page that lays files out as icons or a list, with sorting, folder drill-in, and an adjustable icon size; plus a sidebar of Recent files, Recent searches, and Saved searches.
- Knowledge Agent — Categories and the Knowledge Map are now an evidence-first agent: a proactive review queue, a durable inbox, ad-hoc Q&A, and a live force-directed graph of how your library connects.
- Celebrity / public-figure face recognition — notable faces are auto-named ⭐ with no setup; click a face to jump to every moment they appear.
- People & Faces rebuild — merge/split people, rich per-person appearances, reindex faces from the page, and automatic clustering after a download.
- Media downloader — discover videos on a webpage and batch-download a whole channel or playlist, then transcribe and index them like local media.
- Editable speaker names from the transcript and a speaker library.
- Auto-update check with stable/beta channels and an in-app toast.
- Library management — relink moved/missing media, review failed files and retry just those, remove/forget a file (incl. derivatives), and preview an export before applying it.
Fixed
- Packaged app now bundles ffmpeg/ffprobe, InsightFace, yt-dlp, and resemblyzer so search, faces, the downloader, and diarization all work in the DMG.
Install
Download MediaFind-0.1.4-macos-arm64.dmg, open it, and drag MediaFind to Applications. Signed & notarized by Yongzhe Wang (Team 4428HA4LW3).
sha256 77efa863bd1c2b3860e7f510dbad1a58880494e7edb6e35bad40b33bbf224da8
MediaFind 0.1.3
MediaFind 0.1.3 — macOS (Apple Silicon), signed & notarized.
Fixed
- Search returned "No matches" for every query. The packaged app could not load the local embedding model (sentence-transformers) because the PyInstaller build collected scikit-learn/scipy/transformers data-only, dropping dynamically-imported submodules. It silently fell back to a hashing embedder whose vectors couldn't match the indexed (real-model) vectors, so all searches came back empty. The build now collects the full ML stack and is verified to load the real model + cross-encoder reranker.
If you used 0.1.2 and search came back empty, this release fixes it — your existing library is re-embedded automatically on first launch.
SHA-256: 9f6f8e5cab57bfd262da925b83f1a2ce1b11a984e3dc8048b9fff1c367efdc80
MediaFind 0.1.2
macOS (Apple Silicon) build of MediaFind — signed with Developer ID and Apple-notarized.
What's new in 0.1.2
- Brand-logo detection — find where a brand appears across your media by name ("Nike") or by an uploaded sample logo image (Pro).
- Pro gating for Brand logos, Meetings mode, and People & Faces.
Download: MediaFind-0.1.2-macos-arm64.dmg
SHA-256: 66165c29bd05b32c7d30352f691aca5197c78bc6650c6b2245d952ce5e1a0485
MediaFind 0.1.1
macOS (Apple Silicon) — signed with Developer ID and Apple-notarized.
SHA-256: f95632a99f55e7bd8d28865fa424d5ca6e908acdce26013fc8ce823a869ac662