Frequently asked questions¶

Files and privacy¶

Does ArchiveMind upload my files?¶

No. Indexing reads your files locally and stores what was learned in your local catalogue. Nothing is uploaded, and the beta builds contain no telemetry, analytics, or auto-update. See Privacy.

Does ArchiveMind move, rename, or modify my files?¶

No. Indexing only reads your files. ArchiveMind is a catalogue, not a file manager — it never moves, renames, or edits the originals.

Is ArchiveMind a backup system?¶

No. The catalogue stores information about your files — names, metadata, extracted text, thumbnails — not the files themselves. If a drive fails, ArchiveMind can tell you what was on it, but it cannot restore the files. Keep real backups separately.

What data does ArchiveMind store?¶

Derived data about your files: paths, filenames, sizes, dates, content hashes, metadata, extracted text (up to an extraction limit), thumbnails and media previews, visual-search index data where enabled, repository availability states, and operational logs. Full list: Privacy.

Does ArchiveMind store copies of original files?¶

No — it stores derived data (see above), not the files. Two nuances worth knowing: extracted text can include substantial readable text from inside documents, and thumbnails are small images derived from your photos. Both are stored locally in your catalogue and caches, and both should be treated as personal data.

Who can read my catalogue?¶

Anyone who can use your computer account — and no one else. The catalogue is an ordinary database file on your own disk. ArchiveMind keeps it readable by your user account only, other accounts on the same machine cannot reach it, and full-disk encryption (FileVault on macOS, BitLocker on Windows) protects it if the computer is lost or stolen.

What ArchiveMind does not do is encrypt the file itself. If someone has your logged-in session, or you give them a copy of the file, they can read everything in it — file names, paths, extracted text, and face data if you enabled it. That is the same trust model as your photo library, mail store, or browser profile. Treat the catalogue with the same care as the files it describes: it is a concentrated digest of them, and it stays readable even when the original drives are unplugged.

Why is the download so large?¶

Because the AI ships inside the app. The download (about 2 GB) carries the complete visual-search model and the face-recognition model, plus the runtime that executes them — the app itself is a tiny fraction of it. Most software with these features keeps its download small by running the AI in the cloud, which means your photos do the traveling instead. ArchiveMind makes the opposite trade: the models live inside the app, so visual search and People work on a plane, an unplugged network cable changes nothing, and your files never leave your computer. The size is what the privacy promise weighs.

Does ArchiveMind work without internet access?¶

Yes — indexing, search, browsing, thumbnails, duplicates, and (in the packaged beta builds, which bundle the image-understanding model) visual search all work offline. One documented exception in earlier beta installers: the first launch of the macOS .command-based beta needed the internet once to install its dependencies. Installing optional preview tools such as ffmpeg is a download you choose to make, separate from ArchiveMind itself.

Drives and availability¶

Can ArchiveMind search a disconnected external drive?¶

Yes — this is one of its core features. Once a drive has been catalogued, its contents stay searchable after you unplug it. You just cannot open the original files until the drive is reconnected. See Disconnected drives.

Why can I see a file but not open it?¶

Usually because the drive it lives on is disconnected: search runs against the catalogue, but opening needs the original file. Reconnect the drive. Other causes (the file was moved, renamed, or deleted since indexing) are covered in A file appears in search but will not open.

What does "indexed but unavailable" mean?¶

The file is in the catalogue (indexed), but its original cannot currently be reached (unavailable) — typically because its drive is offline. The catalogued information remains visible; opening the file requires the source to reconnect.

What happens if a file was moved or renamed?¶

The catalogue records the file as it was when last scanned. If it has been moved or renamed since, the catalogue entry points at the old location until you rescan the repository, so opening it may fail. Rescanning updates the catalogue. Details: A file moved or was renamed.

Indexing¶

How long does indexing take?¶

It depends on the number of files and the speed of the drive — minutes for a few thousand photos, potentially hours for multi-terabyte archives on slow drives. You can search while indexing runs. See Indexing.

Can indexing be stopped and resumed?¶

A running scan can be stopped with the Stop Scan button, and stopping is safe — files are only read, and what is already catalogued stays searchable. There is no exact resume-from-checkpoint, but rescans are incremental by default: a new scan recognises unchanged files by size and modification time and skips them, so it spends its time only on what is new or changed. See Indexing.

Why is ArchiveMind still running after indexing finishes?¶

Because the app and the scan are separate things: when a scan finishes, the interface stays available so you can search. ArchiveMind keeps running until you quit it via Quit ArchiveMind. See Reopening ArchiveMind.

Why is the ArchiveMind process using 0% CPU?¶

Usually because it is idle — the scan finished, or it is waiting on a slow drive — not because it is stuck. An open-but-idle ArchiveMind at 0% CPU is normal. See Running but 0% CPU.

Formats and previews¶

What file types are supported?¶

All files are catalogued, hashed, and included in duplicate detection regardless of type. Beyond that, support is by level: photos, documents, audio, video, and archives each get the metadata, text extraction, and preview treatment their format allows. Not every format supports every level, and ArchiveMind does not claim otherwise — where something is not supported, it is labelled honestly rather than faked.

Can every audio or video file be played?¶

No. Playback uses the browser engine that displays ArchiveMind, so it is limited to the formats that engine can decode. A file can be perfectly indexed and still not be playable in the interface — ArchiveMind will say so, and you can open the file in your usual media player instead. See Audio and video.

Why are some thumbnails missing?¶

Common reasons: the format needs an optional tool that is not installed (ffmpeg for waveforms and video frames, rsvg-convert/librsvg for SVG), the format does not support thumbnails, the file's drive is offline, or enrichment has not reached that file yet. ArchiveMind shows an honest placeholder rather than a fake preview. See Thumbnails are missing.

Why are some thumbnails rotated incorrectly?¶

Some cameras store photos unrotated plus an orientation note in the photo's metadata, and software must apply that note when rendering. Current ArchiveMind versions apply the orientation note (including mirrored variants) when generating previews. Thumbnails generated by older versions could come out sideways; ArchiveMind includes a targeted repair that regenerates exactly those previews — originals are never modified. If a thumbnail is rotated while other apps show the photo upright, mention it when reporting a problem.

Search¶

Why are search results sometimes unexpected?¶

Search matches the words in names, paths, metadata, and extracted text — ranked by relevance — so results reflect where your words actually occur, which sometimes surprises. During an active scan, not-yet-extracted files match only by name. See Unexpected search results.

What is visual search?¶

Finding photos by describing what is in them ("beach sunset") instead of by filename. It uses an image-understanding model that, in packaged beta builds, ships inside the app and runs offline. See Visual search.

What happens if a visual-search model cannot load?¶

ArchiveMind says so plainly, and everything else — regular search, browsing, thumbnails, duplicates — keeps working. See Visual search could not start.

Platforms¶

Is there a Windows version?¶

Yes — a private beta for 64-bit Windows. What testers download today is a portable folder: unzip it anywhere and run ArchiveMind.exe.

A per-user installer — no administrator rights, with a Start-menu shortcut and normal entry in Settings → Apps — is now part of the Windows build and will accompany the next Windows beta. Uninstalling never removes your catalogue or settings.

Both are unsigned for now, so the first launch needs one SmartScreen click (More info → Run anyway). See Windows.

Is there an Intel Mac version?¶

Yes — a private beta shipped as ArchiveMind.app, now on beta.6 and level with the Apple-silicon build: built from the same source, with offline licence verification, the application icon, and the Quit that closes the app window. The Intel build is signed with an Apple Developer ID and notarized by Apple, so it opens with a plain double-click. See macOS (Intel).

Intel Macs are supported through 1.0 and for as long as Apple ships security updates for macOS Tahoe (macOS 26), which Apple has said is the last macOS release for Intel Macs.

Is there an Apple-silicon version?¶

Yes — a private beta built natively for Apple silicon, shipped as ArchiveMind-0.1.0-beta.6-macos-arm64.zip containing ArchiveMind.app. It needs no Rosetta 2, unlike running the Intel build on an Apple-silicon Mac. It is the same application as the Intel build, now at the same beta.6 release and likewise notarized — the only difference is the processor each is compiled for. See macOS (Apple silicon).

The catalogue¶

How much disk space does the catalogue use?¶

It grows with the number of files and how much can be extracted from them — extracted text (up to 100,000 characters per file), the search index, and thumbnails are the main contributors. As a real-world reference point: one catalogue of about 500,000 files measured roughly 1.5 GB of database (about 3 KB per file) plus 760 MB of thumbnails (about 113,000 previews at ~7 KB each) — around 2.3 GB in total. Text-heavy and photo-heavy collections sit at the higher end, and visual-search data adds more where enabled. The database lives where you chose during setup, with thumbnails in a thumbs folder beside it — you can check both sizes there at any time.

Can the catalogue be moved to another computer?¶

Yes, with one caveat about previews. The catalogue is a single database file, and it opens fine from a new folder or a new computer — verified by moving a real half-million-file catalogue and searching it from the new location. What carries over and what does not:

Carries over: everything in the database — search, extracted text, metadata, duplicate relationships, and repository records. Drives that are not attached to the new computer simply show as offline, exactly like any disconnected drive, and come back when you attach them.
Does not carry over: thumbnails and previews. They are stored in a thumbs folder beside the database at its original location and are tied to it, so after a move they show as missing at first. They are regenerated from the original files — reconnect the source drives and use the enrichment panel's Generate missing thumbnails operation (or simply rescan).

Getting help¶

How do I report a problem?¶

See Reporting a problem — it lists what to include, and explains the built-in privacy-conscious diagnostics export.