An AI ops workbench that actually runs the commands
SSH, AI agents, and accumulated knowledge in one desktop tool. Ask mode suggests and executes in one click; Agent mode loops through tool calls on its own. Works with Ollama and OpenAI-compatible endpoints, so data never has to leave your network.
What it does
Each capability maps to a concrete everyday problem.
Multi-host SSH
Password and key authentication, grouped sessions, and one-click connect so you stop juggling terminal windows.
Ask mode: conversational diagnosis
Describe the symptom and AI proposes a path with concrete commands. Nothing runs until you approve it.
Agent mode: autonomous loop
AI runs tools in a loop, reads the output, and decides the next step — interruptible whenever judgement is needed.
Skill templates
Turn recurring troubleshooting flows into reusable skills, so one person's experience becomes everyone's starting point.
Local knowledge base
Index runbooks, past tickets, and config conventions so suggestions reflect your actual environment.
Data stays inside
Supports Ollama, OpenAI-compatible endpoints, and DeepSeek — fully usable inside your network.
How it works in practice
Not a feature list — a walk through getting something done.
Describe the symptom, get a path you can run
You do not need to know the exact command upfront. Describe what is broken — service down, disk full, port closed — and AI returns steps in priority order.
Each step carries a runnable command. Nothing executes until you approve it, and the output flows back into the conversation.
From "the service will not start" to the exact line
Agent mode decides what to inspect next: process state, then log tail, then config permissions — narrowing the problem layer by layer.
The loop can be interrupted at any point. If the direction is wrong, stop and redirect without losing context.
Local model plus knowledge base, no traffic out
Point the model at a local or on-prem Ollama instance and every conversation, command output, and diagnostic record stays inside your network.
With the knowledge base enabled, AI prefers your runbooks and past tickets over generic advice.
Design trade-offs
Deterministic work stays in code; judgement stays with you.
Auditable tool calls
Every command execution has a clear origin and recorded output for later review.
Interruptible by design
The agent loop is not a black box — step in and take over at any point.
Credentials stay local
SSH credentials and private keys never leave your machine; model keys stay in local config.
Knowledge that accumulates
Turn a diagnosis into a skill template and reuse it the next time the same problem appears.
What you might want to know first
Does my server data go to the cloud?
Depends on the model you configure. With Ollama or a self-hosted compatible endpoint, nothing leaves your network. With a cloud provider, that provider's policy applies.
What access does it need?
Any account that can SSH in. Use a least-privilege account and keep manual approval on for risky commands.
Which platforms are supported?
Installers for Windows, macOS (Apple Silicon), and Linux (deb).
Start using it now
Free to download, runs locally. Your data stays on your machine.
This is a desktop app — open this page on your computer to download.
Other formats
SHA256 checksum
Provided as-is. The app does not collect usage data.