Other machines & phones
Sync
Hyla's configuration — theme, fonts, keys, agents, Feishu, Tailcat — is one file. Sync keeps it the same on every machine you use, through a server they all can reach: a WebDAV folder, an S3-compatible bucket, or a GitHub repository.
The snapshot is sealed on the machine that writes it. The server stores ciphertext: it cannot read your settings, and neither can anyone who finds the file without your passphrase.
What it syncs
The configuration file alone, the one hyla config path names. It holds what differs from the defaults, so a machine that syncs takes the other machines' settings as they are.
It does not sync what belongs to one machine: sessions and their terminals, projects, the sidebar's SSH and K8s machines, window sizes, or the session you had selected. Those are that machine's own.
Where it goes
| Server | What you fill in |
|---|---|
| WebDAV — Jianguoyun, Nextcloud, ownCloud | the address (as https://dav.jianguoyun.com/dav/), your user and an app password |
| S3-compatible — MinIO, R2, B2, OSS, COS | endpoint, bucket, prefix, region, access key and secret |
| GitHub | owner/name, the path (as hyla/config.enc), the branch, and a token — a private repository, since a snapshot holds your secrets even though they are sealed |
Hyla makes a hyla/ folder in a WebDAV address, as it uses hyla/config.enc in a bucket or a repository, so the same server can hold other things too.
The passphrase
One passphrase, typed on each machine, seals and opens the snapshot. It is derived into a key with Argon2id, and the contents are encrypted with AES-256-GCM: a wrong passphrase, or a file changed on the server, opens nothing at all.
Remember it. There is no recovery: without the passphrase the file cannot be read on any machine, by Hyla or by anyone else.
The passphrase and the server's credentials stay on each machine, in sync.json beside the configuration (readable by your user alone). They are never part of what syncs, so the snapshot does not carry the keys to itself.
When it syncs
- At launch: Hyla pulls, so a machine you have not opened for a while starts with the settings the others have.
- After a change: a setting you change is pushed half a minute later, once the run of changes settles. Turn Sync automatically off to sync only when you say so.
- When you say: Settings › Sync › Sync now, or
hyla syncfrom a terminal.
A machine with no configuration file takes the server's rather than sending an empty one.
When both sides changed
The machines you use do not change the same setting at the same time often, but they can. Then this machine's configuration stays: the server's is written beside it as config.remote-<machine>-<time>, and the push replaces the server's copy. Nothing you wrote is lost without a copy, and Hyla tells you where it put the other side's.
From the CLI
hyla sync # sync now, as the app does at launch
hyla sync status # where it syncs, and how far it got
hyla sync push # send this machine's, whatever the server has
hyla sync pull # take the server's, whatever this machine has
hyla sync set backend webdav # and webdav.url, webdav.user, webdav.password…
hyla sync set passphrase "…" # the same on every machine
hyla sync set auto on
hyla sync set takes backend (webdav, s3, github or off), passphrase, auto, webdav.url|user|password, s3.endpoint|bucket|prefix|region|accessKey|secretKey and github.repo|path|branch|token.