If you live in the rclone, self-hosting, and data-hoarder world, you already know the truth: rclone can move anything, anywhere. Copy, sync, mount, encrypt, bisync - it is the muscle of multi-cloud storage, and nothing else comes close.
But rclone has a deliberate blind spot. It will faithfully move whatever you point it at. What it won't do is tell you which files to move, where your duplicates are hiding, what has gone cold, or what is quietly at risk across half a dozen remotes. That is the gap FileFortress was built to fill.
rclone moves files. FileFortress understands them. FileFortress thinks, rclone acts - they are complements, not competitors. This is what an rclone companion looks like in practice.
rclone is intentionally a transport layer. You tell it source, destination, and flags, and it executes. That focus is exactly why it is so reliable - but it means rclone has no opinion about your data and no memory of it across runs.
For a data hoarder with files spread across Google Drive, OneDrive, an S3 bucket, a Backblaze B2 vault, and a local NAS, the unanswered questions pile up fast:
rclone can answer none of these on its own. You can script rclone lsjson across every remote and stitch the output together yourself, but that is a project, not a tool - and you have to rebuild it every time you want a fresh answer.
FileFortress builds a single local, private metadata index across all of your clouds. It catalogs file names, sizes, paths, timestamps, hashes, and provider-specific metadata into one searchable database that lives on your machine - not in someone else's SaaS.
Crucially, FileFortress does not move, copy, sync, or mount anything. It is not trying to replace rclone, and it would be a worse tool if it tried. It indexes and reasons about your files so that rclone has a plan worth executing.
The metadata index lives on your device, like rclone's own config. FileFortress is privacy-first by design: it reads what your providers expose, stores it locally, and never ships your file listings off to a third party to "analyze" them.
FileFortress currently indexes Google Drive, OneDrive, AWS S3, Backblaze B2, and Local storage. Point it at the same remotes rclone already serves, let it scan, and suddenly every one of those unanswered questions becomes a query.
Once your clouds are in one index, the day-to-day wins are immediate and concrete.
Instead of opening five web UIs to hunt for a file, you ask one question and search every remote together. The index makes cross-cloud lookups instant because the answer is already on your disk.
# Find large files across all connected remotes
filefortress search --filter "size:>1GB"
# Find files untouched in years (cold candidates)
filefortress search --filter "modified:<2021-01-01"This is the one rclone genuinely cannot do for you. FileFortress groups files by name and size, then verifies real duplicates by hash - including copies that live on different remotes entirely. That is how you discover the same dataset paying for storage in three places at once.
# Find duplicate files across every indexed cloud
filefortress find duplicates
# Keep only hash-verified, guaranteed matches
filefortress find duplicates --hash-verified-onlyBecause every file's size and timestamps are indexed, surfacing the heavy hitters and the long-forgotten is a filter, not a forensic exercise. Combine size and age filters and you have a working list of exactly what is worth archiving to cheaper storage or pruning entirely.
Most of your storage bill is driven by a small fraction of your files. Knowing precisely which large, cold, or duplicated files exist - and where - is the difference between guessing at cleanup and acting on it.
Insight is only half the story. Once FileFortress has identified the files worth acting on, it hands the plan straight back to the tool built to execute it: rclone.
The find duplicates command can export its results as ready-to-run rclone commands. Add --export-format rclone and FileFortress applies a keep strategy, then emits the rclone delete commands for everything else.
# Find duplicates, keep the oldest copy, export rclone commands
filefortress find duplicates \
--export-format rclone \
--keep-strategy oldest \
--output-file cleanup.sh
# Review the plan, then let rclone do the moving
# (you run rclone - FileFortress never runs it for you)The workflow is honest about the division of labor: FileFortress writes the plan, you review it, and rclone carries it out. Nothing gets deleted or moved until you run the commands FileFortress generated.
FileFortress does not read your rclone.conf, and it does not run rclone on your behalf. It exports rclone-compatible commands; running them is your call. Always review an exported script before executing it - it is operating on real files.
Because the export is just text, it slots into whatever you already do: pipe it, eyeball it, drop it in a cron job, or paste it line by line. FileFortress hands you a plan in rclone's own language and gets out of the way.
If you already love rclone, FileFortress is not asking you to switch - it is asking to ride shotgun. rclone keeps doing the thing it is unbeatable at, and FileFortress adds the layer of understanding that turns raw transport power into a deliberate strategy.
It is privacy-first the same way rclone is: everything important stays on your machine. The index is local, the analysis is local, and the only thing that leaves is the plan you choose to run.
FileFortress: index your clouds, search across them, find duplicates, surface large and cold files, export a plan. rclone: execute the plan. Two tools, one clean handoff.
--export-format rclone handoffIndex every cloud locally, find the duplicates and cold files rclone can't see, and export a ready-to-run rclone plan. Your data stays yours.