Browse your Linux servers as a file tree and edit files in a real editor with syntax highlighting, find and replace and search across the whole box. Over the SSH access you already have, with nothing installed on the server.
Everything below is what the agent does today, not a roadmap.
Each of these is a specific failure with a specific set of checks behind it, not a general promise to help.
You need to change one line. Instead you are in a terminal editor with no syntax colours, no usable search, and shortcuts you look up every time because you only do this once a month.
What it does. The file opens in a real editor with highlighting, line numbers, find and replace and word wrap, so a one-line change takes seconds and you can see the structure of what you are editing.
The failure mode people actually report with SFTP tools: the save appears to succeed, or the connection dies mid-write, and the file on the server is now empty or half written. On a vhost or a systemd unit, that is an outage.
What it does. Nothing is ever written in place. The new content goes to a temporary file beside the target, is read back and checked against a SHA-256 of what was sent, and only then atomically replaces the original. A failure at any step leaves the original exactly as it was.
A Windows client saving CRLF line endings into a Linux shell script produces the classic "/bin/bash^M: bad interpreter". The file looks fine and will not run.
What it does. Line endings and text encoding are detected when the file opens, preserved on save, and shown in the status bar. A file authored on Linux does not acquire carriage returns because you edited it from a Windows desktop.
A deploy ran, logrotate fired, or another admin was in the same file. A normal editor saves over it and the other change is gone with no trace.
What it does. The timestamp is recorded when the file opens and checked again immediately before writing. If it moved, the save stops and tells you, rather than quietly winning.
The setting is in a config somewhere, you do not remember which, and grep across the box means recalling the flags and reading a wall of output.
What it does. Search a folder or the whole server for a filename or a line of text, and open the match straight in the editor.
The file tree opens like the file browser on your own machine, over the SSH access you already use. Nothing is installed on the server and no port is opened.
Browse the tree, jump straight to a path, or search a folder or the whole server for a filename or a line of text.
A real editor: syntax highlighting, line numbers, find and replace, word wrap. Compare your changes against the version on the server at any point, or revert the buffer if you change your mind.
You press Save, because you are the one making the edit. The previous content is snapshotted first, the write is staged and checksummed before it replaces the original, and the file keeps its own encoding and line endings.
Browsing and reading cost nothing and need no permission, because reading cannot break anything. Editing is different here than everywhere else in the product: you are the author of the change, so there is no approval dialog to click through. The protection is in how the save happens. The previous content is snapshotted, the new content is staged beside the target and verified by checksum before it atomically replaces the original, the file keeps its own permissions, encoding and line endings, and a save stops outright if the file changed on the server while you had it open. When you ask the agent to change something instead, the usual rule applies: it shows you the command or the diff and waits.
Read the security modelInspect containers, read logs, check stats, restart services and run Compose stacks, just by describing what you want.
Read more →Test configs before they go live, read the right log, and reload safely, on nginx, IIS or Tomcat.
Read more →Run your existing deploy scripts, watch the output live, and get a straight answer about whether it worked.
Read more →Read-only checks and queries over your existing SSH connection, with credentials encrypted at rest.
Read more →Issue trusted, auto-renewing certificates, wildcards included, and get the paths handed straight back.
Read more →Read EC2, S3, Elastic IPs and EBS, spot what is quietly costing you money, and make changes only after you approve them.
Read more →Point it at one server. You will know within an hour whether it earns its place.