- Target is reachable from the public internet (no VPN, no IP allow-list against TigerGate’s egress IPs).
- You’re OK with the scan traffic coming from TigerGate’s IP block (we publish them on request for allow-listing).
- Target is on a private subnet, behind a VPN, or under an allow-list TigerGate isn’t on.
- You need to scan a private codebase that you can’t push to a public Git host.
- You have a corporate policy forbidding cloud SaaS from probing your infrastructure.
Run your first hosted scan
Configure an LLM provider
Dashboard → Pentest → Settings → AI Providers → Add provider.Pick OpenAI, Anthropic, Google Gemini, Azure OpenAI, or any OpenAI-compatible endpoint. Paste your key, click Test — the dashboard makes a tiny live call to confirm.Provider keys are stored encrypted at rest in TigerGate. They’re injected into each scan job at runtime. See LLM providers for the full list.
You can run TigerStrike without configuring a key, but in that mode the scan only runs the deterministic-checks portion (no agentic exploitation). For real pentest results, configure a provider.
Set a budget (recommended)
Dashboard → Pentest → Settings → Budgets.
- Per-scan budget (USD) — hard cap for one scan. Default: $25.
- Monthly org budget (USD) — hard cap across all scans this month. Default: unlimited.
budget_exhausted, and the dashboard shows the cost so far. See Budgets & spend.Create a target
Dashboard → Pentest → New Scan.
| Field | Notes |
|---|---|
| Target type | Web app / API / Codebase |
| Target URL | https://staging.example.com (must be publicly reachable) |
| Scope | auth, api, auth+api, or full |
| Max duration | Wall-clock cap, in minutes (default 60) |
| Per-scan budget | Defaults to your org-level setting; you can lower it for this run |
| Runner | TigerGate hosted |
Start scan
Click Start scan. The scan moves through
queued → running → completed (or budget_exhausted / cancelled / failed).Authenticating against the target
Most real-world targets need auth. TigerStrike supports:| Auth | How to configure |
|---|---|
| None (anonymous) | Default |
| Basic auth | Username + password fields in the scan config |
| Bearer token | Static token; pasted into the scan config |
| Form login | Provide login URL + username + password fields. The agent navigates the form on each session. |
| OAuth (PKCE) | Provide client ID, scopes, the agent walks the flow |
| Custom session cookie | Paste cookie value(s); useful for quick sandbox logins |
Scope
Picking the right scope matters for both quality and cost:auth— only the auth surface. Login form, password reset, MFA, session management. Cheapest scan.api— only API endpoints (REST/GraphQL/SOAP). Skips the login flow; assumes you’ve provided a token.auth+api— login first, then probe everything you can reach as the authenticated user. Good general-purpose scope.full— everything above plus deep crawl. Most thorough; most expensive.
Reusable scan configs
After you run a few one-off scans, save the common settings under Pentest → Scan Configs and reuse:- One-off — pick Pentest → New Scan and select a saved config.
- Scheduled — schedule a config to run nightly / weekly under Pentest → Schedules.
- CI-triggered —
POST /v1/pentest/scanswith the config ID. See the API reference.