What we store, what we don't, and why.

The short version

OperationLeadAI is a hosted service: creating an account and running a scan involves sending data to our server, unlike an earlier local-only version of this tool. Your account (email, hashed password, company name) and every scan you run — both the computed results and the job/estimate/invoice files themselves — are stored in an encrypted, automatically-backed-up database, so nothing you've uploaded or found is ever lost. Payment details never touch our server at all — Stripe handles those directly.

Account data

When you sign up we store your email, company name, dashboard style preference, and a hashed password (via scrypt, a one-way function — we cannot read your actual password, and it is never logged or stored in plain text). Signing in checks your password against that hash and issues a session cookie so the server knows you're you.

Sign-in activity

We keep a short log of sign-in attempts against your account — timestamp, success or failure, IP address, and browser identifier — so suspicious activity (repeated failed attempts, an unfamiliar location) can be spotted. This is a security measure, not analytics; it's not used for tracking or marketing, and never shared.

Resetting your password

Requesting a reset generates a random, one-time token (not derived from your password) that expires in 1 hour and is emailed to the address on file. We respond the same way whether or not that email has an account, so this can't be used to check who has signed up. Completing a reset signs out every other active session on your account, in case the reset was needed because someone else had access.

Job, estimate, and invoice data

Running a scan sends the files you provide to our server so the detection engine can process them, and — so nothing about a scan can ever be lost, misplaced, or need re-uploading — both the raw files themselves and the computed results (risk score, total exposure, findings) are saved to your account. This is stored in a database with encryption at rest and automated backups, access-scoped so only your own signed-in session can read it back. Demo scans (the built-in sample dataset) are never saved, since they aren't your business's real data. If you download a report (HTML, CSV, or JSON), that file is generated on the fly and streamed to you.

Scan history

Every real scan you run — files and results together — is kept so you can look back, re-download an old report, or track whether things are actually improving over time. You can delete your entire scan history at any time by asking us to delete your account.

Payment data

We use Stripe to process upgrades. Card numbers and billing details go directly to Stripe — our server never sees, transmits, or stores full card numbers. We store only what Stripe tells us back: a customer reference ID and your subscription status, so we know whether your account is on the free trial or upgraded.

Cookies

One: a session cookie (httpOnly, so page scripts can't read it) that keeps you signed in. It's functional, not tracking — we don't run analytics, ad pixels, or third-party trackers.

Optional AI features

If a Groq API key is configured on this deployment, two features send a limited amount of data to Groq, a third-party AI provider, to do their job:

• Executive summary. Sends the deterministic findings already calculated — risk type, dollar amounts, confidence, and sample size. Never sends your raw job, estimate, or invoice rows.
• Smart column/file matching. Only when your file's columns or file type can't be confidently matched automatically: for ambiguous file types, up to 3 sample rows (every column in that file) may be sent; for individual ambiguous columns, just sample values from those specific columns are sent — not the whole file. This only happens for genuinely unclear cases; most exports are matched without ever calling AI.

Without a configured key, every feature still works — the deterministic risk engine never depends on AI, and file matching falls back to asking you to assign roles manually. See Groq's own privacy policy for how they handle data sent to their API.

Stellar, the AI assistant

Stellar is a chat assistant (also powered by Groq) that answers questions about your scan. When you send Stellar a message, we send along: your message and recent chat history in that conversation, your current scan's computed results (score, exposure, findings — never raw file rows), and a short summary of your recent scan history (dates, scores, exposure totals) so Stellar can speak to trends. Stellar is instructed to ground every answer in that data, name where a figure came from, and ask you a clarifying question rather than guess when your request is ambiguous.

Your conversation with Stellar is saved to your account so it's there the next time you sign in — it's stored the same way your scan history is, tied only to your account, never shared or used to train anything. You can delete it at any time by asking us to delete your account.

Your rights and data retention

Your account and everything tied to it — scan history, uploaded files, Stellar conversations — are retained until you ask us to delete them. Email us and we'll delete your account and everything in it, including from backups within their normal retention window. Until then, it's kept durably (encrypted, backed up daily with point-in-time recovery) specifically so it doesn't get lost by accident.

Changes to this policy

If what this service does with data changes, this page will change to describe it accurately — the commitment is to match what the software actually does, not the other way around.

Contact

Questions about this policy, or to request account deletion: hello@operationleadai.beaumiertechnologies.com.