File transfer compatibility matrix
Short answer
Seven ways to move a file, checked against the ten situations that actually decide which one works. Read down a column to find the methods that survive your situation.
The short version: no method wins everywhere. Cables and local tools win on speed and privacy but need proximity or a permissive network. Internet-based methods survive restrictions but cost upload time and involve a third party.
The matrix
Works means it reliably does the job. Depends means it works with a caveat, always stated. Fails means it does not work in that situation at all. Unverified would mean we could neither test it nor find it documented — currently nothing is marked that way.
| Method | iPhone → Mac | iPhone → Windows | Android → Windows | Same normal Wi-Fi | Isolated campus Wi-Fi | Different networks | Receiver offline now | No install allowed | Large files (>1 GB) | Sensitive files |
|---|---|---|---|---|---|---|---|---|---|---|
| AirDrop | Works | FailsApple devices only | FailsApple devices only | Works | Workspeer-to-peer — never uses the Wi-Fi network | Fails | Failsboth devices must be present | Worksbuilt into iOS and macOS | Works | Worksnever leaves the two devices |
| USB cable | Works | DependsWindows needs Apple drivers; HEIC conversion | Works | Worksno network used | Worksno network used | Worksno network used | Failsneeds both devices in hand | Dependsdrivers may be required on Windows | Worksfastest option | Works |
| LocalSend | Works | Works | Works | Works | Failsclient isolation blocks peer discovery | Failssame local network only, by design | Failsboth devices must be online together | Failsinstalls an app on both devices | Works | Worksnever leaves the local network |
| PairDrop | Works | Works | Works | Works | Dependslocal discovery fails; code pairing still works | Workspaired devices connect regardless of network | Failsboth devices must be online together | Worksruns in the browser | Works | Dependspeer-to-peer, but relayed via TURN when needed |
| Cloud storage | Works | Works | Works | Works | Worksordinary outbound HTTPS | Works | Worksfile waits in the account | Worksweb interface | Dependsbounded by your storage plan | Dependsprovider can typically read consumer files |
| Email to yourself | Works | Works | Works | Works | Works | Works | Works | Works | Failstypically capped at 20–25 MB | Failsstored by your provider indefinitely |
| Messaging apps | Works | Works | Works | Works | Works | Works | Works | Worksweb clients available | Dependscaps vary; media is often recompressed | Dependsvaries by app; several are end-to-end encrypted |
| todevice | Works | Works | Works | Worksworks, but a local tool will be faster | Workstested on an isolated network | Works | Worksheld encrypted for up to 1 hour, then purged | Worksruns in the browser | Fails500 MB per file during the beta | Dependsencrypted in-browser; see the privacy page for the limits |
How these were determined
Platform behaviour (AirDrop's Apple-only restriction, mail attachment caps, local discovery requirements) comes from vendor documentation and from how these protocols work. The todevice row comes from our own testing, including a 480 MB transfer through production and a transfer across an isolated network.
We have an obvious interest in this table, which is exactly why todevice is marked Fails on large files and Depends on sensitive files rather than being given a clean sweep. A comparison where the author wins every row is not a comparison.
Dates on this page are real: it was published and last checked on the dates shown at the top. If something here is now wrong, we would rather hear about it than leave it.
Reusing this table
You are welcome to reference or reproduce this matrix — in a student handbook, a help-desk article, a comparison post — with attribution and a link back to this page. Please keep the caveats attached to the cells; they are what make the table honest rather than a scoreboard.
If a row is wrong or a method is missing, tell us and we will fix it and note the change.