You authorize the work
PenThreatAI is for authorized security testing. You decide the target, the scope, and when local tools may run.
PenThreatAIPenThreatAI
Security you can read.
How we process data, where agent code runs, and how billing works — then start free when you're ready.
See how the platform worksWhat we commit to.
Four non-negotiables before the vendor list. Authorized work, deletion you control, payments that never touch our servers, and a commercial product — not a bait-and-switch repo.
You control deletion
Delete tasks, sandboxes, shared links, or your whole account from Settings. Canceling removes paid access at period end.
Cards stay with Stripe
Payment details never hit our servers. Billing, invoices, and portal access live under Settings → Account.
No open-source bait
PenThreatAI is a commercial product. This page is the public record of how we process data — not a public repo dump.
How it works
Authorize. Process. Delete.
One engagement loop. You stay in control of scope and what remains after the work is done.
- 01
You scope the engagement
Authorized target, prompts, uploads, and whether cloud or local tools may run — all your call.
- 02
We process what the task needs
Messages, files, sandbox output, and search queries go to the providers required to run the agent.
- 03
You delete when you're done
Wipe tasks, sandboxes, shares, or the account. Paid access ends with the billing period when you cancel.
Inventory
Everything we process.
Depending on how you use PenThreatAI, the service may process the following.
- Prompts and task messages you send to the agent
- Files you upload (PDF, CSV, JSON, TXT, Markdown, DOC/DOCX, and PNG/JPEG/WebP/GIF), including extracted text
- URLs and search queries when the agent browses or searches the web
- Terminal commands and output from agent sandbox sessions
- Browser screenshots captured inside the agent sandbox
- Notes and custom instructions you save
- Account information such as email and name
- Usage analytics and error diagnostics
AI model providers
Task requests route through OpenRouter to the provider behind the model you select — Anthropic, Google, DeepSeek, Moonshot AI, xAI, and others. Your prompt, relevant conversation context, and tool results go to that provider to generate a response.
Web search uses Perplexity; page retrieval uses Jina AI. Message content is screened by OpenAI for moderation. Training use varies by provider — see Subprocessors, not a blanket guarantee from us.
Execution environments
By default, terminal and browser actions run in an isolated E2B cloud sandbox. Delete sandboxes anytime from Settings → Data controls.
Local execution runs on your machine with your privileges and no isolation — use it only on machines dedicated to testing. Setup lives on Download.
Account security
- WorkOS AuthKit handles authentication — we don't operate a password database
- TOTP MFA available in Settings → Security
- Revoke sessions per device or everywhere from Settings → Security
Encryption & transport
- Web app and APIs over HTTPS/TLS in transit
- Uploaded files in encrypted AWS S3 storage
- Sessions issued and validated by WorkOS
- Payment traffic stays with Stripe Checkout and the customer portal
Regions
Where your data is processed.
Placement follows where you connect from — not an account toggle.
Agent runs, new cloud sandboxes, and new file uploads place by connection location. Europe uses our EU region (Frankfurt, eu-central-1). North America uses the nearest US region. A VPN or travel changes it.
Sandboxes and files created before regional placement stay where they were until you delete them. The primary database, AI providers, search, moderation, auth, billing, and analytics are not regional — see the Privacy Policy for international processing.
Billing & Extra Usage
Stripe processes payments; card details never reach our servers. Manage or cancel in Settings → Account. Deleting your account cancels any active subscription.
Extra Usage is prepaid with a monthly cap you control. Flip it on in Settings → Extra Usage when a plan budget runs out — charges draw from your balance, not an open tab.
Data deletion
- Delete individual tasks, or all tasks at once
- Delete your terminal sandboxes
- Revoke links to tasks you have shared
- Delete your account — removes database records, cancels Stripe, and deletes your auth record
- Recreating an account does not reset usage limits or referral eligibility
Vendors
Subprocessors.
These services process data on our behalf to run PenThreatAI.
| Provider | Purpose | Data categories |
|---|---|---|
| OpenRouter | Routes task requests to third-party AI model providers | Prompts, conversation context, file content, tool output |
| OpenAI | Content moderation | Message content |
| Perplexity | Web search performed by the agent | Search queries from agent runs |
| Jina AI | Webpage content retrieval for the agent | URLs and retrieved page content |
| E2B | Cloud sandboxes (US, or EU for connections from Europe) | Commands, command output, files inside the sandbox |
| Convex | Primary application database | Account data, tasks, messages, files, settings |
| Amazon Web Services (S3) | File storage (US regions, or eu-central-1 for connections from Europe) | Uploaded files |
| Upstash / Redis | Rate limiting and response stream resumption | Transient identifiers and streaming state |
| WorkOS | Authentication and account management | Email, name, sessions, MFA enrollment |
| Stripe | Payments and subscription billing | Billing contact and payment details (held by Stripe) |
| PostHog | Product analytics and error tracking | Usage events and diagnostic data |
| Trigger.dev | Background execution of long-running agent tasks (US regions, or eu-central-1 for connections from Europe) | Task payloads including conversation context |
| Vercel | Application hosting | Request data processed by the web application |
| Intercom | In-app customer support messaging | Account identity and support conversation content |
Reports
Responsible disclosure
Found a security issue in PenThreatAI? Report it through in-app support chat. We review good-faith reports. We don't run a paid bug bounty program at this time.
Assurance
Compliance
PenThreatAI doesn't currently hold SOC 2, ISO 27001, or other third-party certifications. The service is offered as described in our Privacy Policy and Terms of Service. We'll update this page as our compliance program develops.
Subscribe
PenThreatAI Ultra
Trust the stack. Free to try. Ultra from $200/mo when you need headroom — no card required to start.