Open prod from a GUI group or from any shell.
Profiles, credentials, groups, ProxyJump bastions, and port forwards live in one local database. The GUI is for browsing and editing; the CLI is for muscle memory.
Terminal sessions, credentials, tunnels, files, recordings, and AI diagnosis in one local-first cockpit.
Color-coded command blocks make long terminal history readable. The desktop app and bundled CLI share the same SQLite database. Secrets stay in your platform keychain; encrypted sync goes to your own private GitHub repo.
RSSH keeps the ordinary SSH loop intact, then adds structure around it: connection profiles, readable command blocks, explainable AI probes, replayable sessions, and sync that does not require a vendor cloud.
Profiles, credentials, groups, ProxyJump bastions, and port forwards live in one local database. The GUI is for browsing and editing; the CLI is for muscle memory.
RSSH uses local xterm markers, not fragile prompt parsing. Each command block gets a golden-angle color, so thousands of lines still show where work started, failed, retried, and recovered.
The assistant can choose a Java, Go, CPU, memory, IO, or network playbook, but every command passes shape validation and lands on a human approval card before it runs.
java -version && uname -a
Read-only probe. No files changed.
Shell completions are generated for zsh, bash, fish, and PowerShell. The CLI can open sessions, manage config, launch forwards, and push encrypted backups.
rwsyncrwPrivate keys and passwords stay in OS keychain storage. Each credential decides whether it can sync remotely, and the remote target is your private GitHub repo.
Tell RSSH what is wrong. It picks the playbook, probes the environment, samples lightly, and attributes the likely cause. You never type jstack, pprof, or jmap by hand just to start the investigation.
Java, Go, CPU, memory, IO, network, and shell context are selected from the symptoms.
Destructive forms are rejected before a command reaches the approval card.
The assistant gathers and explains evidence. It does not apply fixes on your server.
User intent stays plain. The assistant turns it into an investigation plan.
RSSH samples environment facts before choosing Java, Go, process, network, or disk tools.
java -version && uname -a
Read-only probe. No files changed.
Large files and samples are pulled back for local analysis instead of being pasted into the model context.
The final answer ties symptoms to evidence and leaves the operator in control of the fix.
Every command becomes a block. A colored edge marks its output until the next command begins. The color sequence is generated by golden-angle HSL rotation, giving stable contrast without remote hooks or shell prompt hacks.
The bundled CLI reads the same SQLite database as the desktop app. Create a profile in the GUI, then run rssh open prod from the terminal. Shell completions cover zsh, bash, fish, and PowerShell.
editrwplaylivereadsshfwdsyncSecrets live in your platform keychain. You decide per credential what can sync. The remote destination is your own private GitHub repo, and payloads are encrypted before upload.
save_to_remote decides what can enter an encrypted backup.
Private keys and passwords are stored in platform secret storage. Android falls back to encrypted SQLite for the mobile environment.
Production root key? Keep it local. Dev password? Allow it to sync. The decision is attached to each credential instead of a global all-or-nothing switch.
The sync target is a private GitHub repo you own. RSSH does not need accounts, hosted config storage, or telemetry to keep devices aligned.
AI diagnosis also follows the security model: commands are validated, explained, and explicitly approved before execution.
RSSH includes the operational details that usually force teams to juggle terminal tabs, SFTP apps, tunnel scripts, shell history, and scattered SSH config files.
Remote file browsing with upload and download, attached to the same SSH profiles you already use for terminal sessions.
Local, remote, and SOCKS5 forwards with named configs and live traffic stats.
Record and replay sessions as asciicast v2 with speed control.
Highlight ERROR, WARN, INFO, or your own terms inside long scrollback.
~/.ssh/configBring existing SSH config into the profile manager instead of recreating every host.
Reusable commands open with Cmd+E, and broadcast can send input across multiple sessions when you are coordinating repeated work.
Ctrl, Alt, Esc, Tab, and arrows are reachable on Android with safe-area-aware navigation.
Open zsh, bash, or PowerShell locally when the task starts outside SSH.
Run RSSH inside a JetBrains IDE tool window while sharing the same ~/.rssh data directory.
Go deeper into the decisions behind RSSH: why the terminal is a natural ops surface, how AI diagnosis stays operator-approved, and how local-first sync keeps credentials under your control.
The product philosophy behind RSSH and why the SSH client is the right place for operational context.
AIHow routing, approval, shape validation, and local analysis fit together.
BlocksCommand block markers, golden-angle colors, and keyword highlighting.
SecurityKeychain storage, per-credential remote flags, and encrypted backup boundaries.
WorkbenchSFTP, forwards, recording, snippets, search, and mobile terminal ergonomics.
Releases are published on GitHub. The cards below resolve to the newest matching asset when the GitHub API is reachable, and fall back to the latest release page otherwise.