← All documentation
DOCS · GETTING STARTED

Getting started

Sign up, send your first prompt, register a server. Five minutes.

01

Create an account

Head to /login and register with email or an OAuth provider. You start on the free tier with monthly credits — no card required.

Once you’re in, you land on the dashboard at /dashboard. From there you can open the chat, manage projects, register SSH servers, and watch your credit usage.

02

Send your first prompt

Open the chat at /chat. You can either type a free-form question or use a slash command. Try one of these to feel the difference between generic chat and a router-driven security platform:

prompt
copy
/scan acme.com
/checkleak [email protected]
/cve CVE-2024-3094
/sigma LSASS access via WinAPI

Each command activates the matching skill and runs real lookups in parallel. The output streams in block-by-block: short prose, structured findings cards, and code snippets where relevant.

Tip
Press / in the composer to open the slash-command picker. Press to recall your last message and edit it.
03

Spin up a project

For anything bigger than a one-shot question, create a project. Open /dashboard/projects New project. Add:

  • Assets — domains, IPs, apps, services, servers in scope.
  • Tech stack tags — what the target runs (so the model picks up the right CVEs and TTPs).
  • Goals— what “done” looks like.
  • Notes — markdown engagement journal.

Every chat you start inside the project picks up all of that as context — no re-pasting scope each turn.

04

Register an SSH target

To execute commands on your own infrastructure, register a target at /dashboard/servers. Provide the host, username, and an SSH private key (PEM). Click Test to verify reachability — we run a one-shot whoami.

From any chat, mention the server with @<name> or fire /run @<name>. The model drafts the command, you click Run, the platform executes over SSH and streams output back into the chat.

Heads up
The private key is stored AES-256-GCM encrypted; it’s never returned to the client. See Security & data for the full model.
05

What to read next