Skip to content
ANTHRIONANTHRION
Launch App

Your first scan

This is the end-to-end walkthrough for the dashboard: sign in, start a scan, watch it run live, and read the report. It takes a few minutes and, during launch, costs nothing. If you want to drive scans from code instead, jump to the REST API or the MCP server.

1. Sign in

Open app.anthrion.xyz and choose CONNECT.

The ANTHRION app sign-in screen: the ANTHRION wordmark above a single magenta CONNECT button.
The sign-in screen at app.anthrion.xyz — one button to get in.

Sign in with an email or a wallet — whichever you prefer:

  • Email — enter your address and we send a one-time code to your inbox. Paste the code and you are in. There is no password to manage.
  • Wallet — connect a wallet and approve the sign-in. We never ask for a private key or seed phrase, and ANTHRION takes no custody of your funds.
The sign-in dialog offering email and wallet options.
Sign in with an email one-time code or by connecting a wallet.

2. Get your bearings — the Overview

After signing in you land on the Overview. It is your account at a glance: findings grouped by severity, scans grouped by status, and your most recent activity. The left sidebar is how you move between scans, the endpoint registry, and your settings.

The ANTHRION Overview dashboard showing findings by severity, scans by status, and recent activity, with the left navigation sidebar.
The Overview — severity counts, scan statuses, and recent activity. This is the one place outside a report where severity colours appear.

3. Start a scan

Choose New scan (from the sidebar or the Overview). Pick a scan type, then fill in its target.

The New scan form with the scan-type picker showing nine options and the AI / LLM attack target fields.
Pick one of nine scan types, then give it a target — a live endpoint, a public repo, or pasted text.

The scan type decides what ANTHRION inspects and what kind of target it expects:

| You want to test | Pick | Target you provide | |---|---|---| | An AI agent / LLM | AI / LLM attack | A live agent endpoint, or a pasted system prompt | | A live website | Web app vuln | A site URL (single page or a multi-page crawl) | | An API | API security | An endpoint URL, or a pasted OpenAPI/Swagger spec | | A Web3 dApp | Web3 dApp | A dApp URL + chain (read-only — never signs) | | Code for secrets | Leaked secrets | A public GitHub repo, or pasted code | | Copied / forked code | Code similarity | A public GitHub repo | | A repo's trustworthiness | GitHub trust score | A public GitHub repo | | A paid (x402) endpoint | x402 endpoint verify | An endpoint URL (read-only — never pays) | | Source code, statically | White-box (SAST) | A public GitHub repo, or pasted code |

See Scan types for exactly what each one detects. When the target is set, choose Start scan.

Safe targets

Only scan something you own or are authorised to test. For trying things out, use a target built for it — for example the public, intentionally-vulnerable test site testphp.vulnweb.com, or one of your own staging URLs or public repositories.

4. Watch it run, live

You are taken straight to the scan, where progress streams live. The status moves through the lifecycle — QUEUED, then RUNNING — and findings appear the moment they land, so a high-severity issue surfaces early rather than only at the end.

A running scan showing a RUNNING status chip and a live progress stream of the phases the scan is working through.
A scan in progress — phases stream in as the engine works, and findings appear as they are confirmed.

You do not have to wait at the page — the scan keeps running, and you can come back to it from Scans at any time. More on this in Real-time results.

5. Read the report

When the scan reaches DONE, the page becomes the report: a severity summary at the top, then every finding in full. Each finding follows the same normalized model — so a finding from an AI scan and one from a code scan read the same way.

A finished scan report: a severity summary followed by a list of findings, each with a severity badge and a title.
A finished report — a severity summary, then the findings, each severity-ranked.

Open a finding to see the three things you act on:

  • Severity — how urgent it is, from CRITICAL down to INFO.
  • Evidence — the supporting detail: the exchange, request/response, or code path behind the finding, so you can confirm it for yourself.
  • Recommendation — concrete guidance on how to fix it.
A single finding expanded, showing its severity badge, description, an evidence block, and a recommendation.
A single finding: severity, the evidence behind it, and a recommendation you can act on.

6. Download the report

Use Download report to save the report as a PDF — the same findings, evidence and recommendations in a shareable file you can hand to a teammate or attach to a ticket. Reports are private to your account.

Honesty

A scan is an indicator of risk, not a guarantee of security. A clean result means ANTHRION's checks did not surface a finding — not that the target is provably safe. Where a scan could only partially cover a target, the report says so rather than overclaiming.

What it costs

You pay per scan, in USDC — and during launch, scans are free. There is one free scan per wallet to start, no subscription, and no custody of your funds. Full pricing and on-chain payment details land under Pricing and Payments as they are finalized.

Next steps