Remote Files (SSH/SFTP)

Orbitron opens files on a remote host (e.g. an HPC login node) directly over your system ssh, with no server to deploy. A remote target is a scp-style spec [user@]host:/remote/root: the segment before the first : is the SSH destination and the rest is the directory that relative request paths resolve against. Connection details — port, identity file, ProxyJump, 2FA — come from your ~/.ssh/config and ssh-agent, exactly as a normal ssh/scp would. A requested file is fetched whole into a temporary location and parsed with the same loaders used for local files, so remote format support is identical.

The remote target is shared across the frontends: - Desktop: set it in Preferences → Remote data, then browse with File → Open Remote…. - CLI: pass --remote [user@]host:/remote/root (or persist it in the config). - Python: orbitron.Orbitron(remote="[user@]host:/remote/root").

The ORBITRON_REMOTE environment variable overrides the saved target for a single process.

6.1 Typical HPC workflow

Confirm you can reach the host with plain ssh. For a non-default port, identity, or a jump host, add a Host alias to ~/.ssh/config — Orbitron uses it automatically:
```
Host mycluster
  HostName login.cluster.example.edu
  User myuser
  ProxyJump bastion.example.edu
```
Point Orbitron at the remote root (the alias stands in for [user@]host):
- Desktop: Preferences → Remote data, enter mycluster:/scratch/myuser/runs, click Test Connection, then Save.
- CLI: orbitron --remote mycluster:/scratch/myuser/runs info benzene.out
- Python: orbitron.Orbitron(remote="mycluster:/scratch/myuser/runs")
Browse and open files. The SSH session opens on first access — so any key passphrase or 2FA prompt appears then — and is reused for the rest of the session. The saved target persists, so future runs reconnect automatically.

Nothing needs to be installed or kept running on the remote host beyond the standard sftp-server subsystem that OpenSSH ships with. No tunnel, port, or bearer token is involved.