SEC 00 // SITREP

Your AI agent,
on every server.

sgt.sh is SSH built for agents. Deploy an agent, attach it to a single server or your entire fleet, and point Claude Code, claude.ai, or Cursor at one MCP URL. The agent gets hands; you keep command — it never holds a key, dangerous commands wait for your approval, and everything it does can be replayed.

Deploy your first agent See the 60-second setup
  • One MCP URL, whole fleet
  • No keys on the agent
  • Human approval on risky commands
  • One-click kill switch

SEC 01 // AGENTIC SSH

Give agents hands. Keep command.

AI agents are already writing your code and triaging your incidents. sgt.sh is how they touch real servers: scoped, supervised, disposable access — never a raw key in a config file.

One URL. Any agent. Every server.

Point Claude Code, claude.ai, Cursor — anything that speaks MCP — at a single hosted URL. Your agent lists its servers and runs commands immediately. No SSH config, no jump hosts, no keys to wire into yet another tool.

Attach one server — or the whole fleet

An agent's reach is a list you control. Start it on a single staging box; when it earns trust, attach the web tier, or everything you own. Detach an endpoint and the agent loses it mid-conversation — scope is re-checked on every single command.

Guardrails, not vibes

Every command the agent runs passes your org's policy engine. Allow the routine, block the reckless, and route the gray area — restarts, deletes, schema changes — to a human approval queue. The agent waits politely; you decide in one click.

A kill switch that means it

Revoke an agent and it's over: the token dies, its live sessions are torn down mid-command, and nothing it ever held lets it back in — because it never held anything. No credentials to rotate, no laptops to wipe, no doubt.

Replay every move

Person or agent, every session lands in a tamper-evident audit log. See exactly what your agent did at 3 a.m., prove it to your auditors, and settle "who changed that?" with a replay instead of an argument.

Humans welcome too

The same fleet is yours from any browser: full terminal, file transfer, nothing to install. Grant a teammate access by name, revoke it in a click — the humans get the same clean, keyless treatment the agents do.

SEC 02 // CONNECT

From zero to agent in 60 seconds

Deploy an agent in the dashboard, pick which servers it may touch, and copy one config line. That's the whole integration: your agent token is the only secret, and it can't do anything your scope and policy don't allow.

Prefer batteries included? The built-in AI chat lets you bring your own Anthropic, Gemini, xAI, or OpenAI key and work your fleet from the dashboard — through the exact same scoped, recorded path. No side channels, no unlogged shells.

Deploy an agent
.mcp.json 
{
  "mcpServers": {
    "sgt": {
      "type": "http",
      "url": "https://sgt.sh/mcp",
      "headers": {
        "Authorization": "Bearer ${SGT_AGENT_TOKEN}"
      }
    }
  }
}
tools: list_endpoints · run_command · wait_for_approval scoped to the servers you attach

SEC 03 // UNDER THE HOOD

Why it's safe to hand an AI a shell

Not because we promise the agent will behave — because the plumbing makes misbehaving pointless. The boring, bank-grade details, so you don't have to think about them:

  • Agents never hold keys. Your SSH keys stay sealed in our vault. Each session gets a certificate signed by your org's own CA that expires in five minutes and only works from our relay — steal it and it's already dead.
  • One session, one server. Every session runs in its own sealed sandbox that can reach exactly the machine it opened — an agent gone sideways has nowhere to go.
  • Close port 22 to the internet. All sessions arrive from our published relay IPs (or your own dedicated IP). Allowlist those, and your servers go dark to everyone else on the planet.
  • Rotate nothing, revoke anything. Certificates expire on their own; agent and team access dies the moment you click revoke — live sessions included.
Add your first server

SEC 04 // PRICING

Rules of engagement

One plan, everything included. You pay per server, not per seat — deploy as many agents and add as many teammates as you want, free.

Base Deployment

Pack

$9.95/mo

  • 10 servers included
  • Unlimited AI agents & teammates
  • Hosted MCP + AI chat
  • Policy engine & approval queue
  • Web terminal & SFTP
  • Audit trail & recordings
Deploy the Pack

Grows With Your Fleet

Scale

+$1.00/mo per server

  • Beyond the first 10 servers
  • Same agents, policy & audit — nothing locked
  • Add or remove servers anytime
  • Prorated to the day
  • 50 servers ≈ $49.95/mo, all in
Scale Up

Hardened Egress

Dedicated IP

+$4.95/mo each

  • One egress IP, one tenant: you
  • Pin firewall rules to a single address
  • Never shared, never recycled
  • Stack as many as you need
Claim an IP

SEC 05 // ORDERS

Give your agent its orders.

Add a server, deploy an agent, and watch it run its first command in under five minutes.

Get started — $9.95/mo