Security & data handling, stated plainly.
This page describes how we actually handle client data and access client systems — including what happens every time we log in remotely — written to match our real practices, not a template.
Core principles
- Least access. We request access only to the systems and data a specific workflow requires — scoped, documented and approved by you before use.
- Official interfaces only. We integrate through vendor-supplied APIs, widgets and hosted pages — not direct database connections or scraped workarounds.
- Human approval on high-risk actions. Customer-facing communications, payments and anything compliance-sensitive keep a human in the loop by design.
- Your data stays yours. We don't retain client data after an engagement ends beyond what's needed for invoicing and legal obligations.
- No secrets in code. API keys and credentials live in server-side environment variables and secret managers — never in client-side code, repositories or logs.
Data handling in practice
What we collect
Forms on this site collect only what's needed to respond to you: name, work email, company, phone and your message. We never ask for patient records, ID documents or financial account details through web forms.
Storage & transport
All traffic to our services is HTTPS-encrypted. Submission data is stored with minimal-privilege access controls on managed cloud infrastructure (hosted in Singapore by Render; see our Privacy Policy for the full list of locations and recipients). Form submissions are rate-limited and bot-protected.
E-signature documents
Agreement PDFs and signature images are stored on encrypted persistent storage. Each signer receives a unique access token; every view, share and signature is event-logged for audit. Documents are only accessible to their signers and authorised administrators.
Payments
Payments run through Stripe's hosted checkout. Card data never touches our servers — we only ever see a payment status and a session identifier.
How we access your systems
When we work on your systems remotely, three safeguards apply by default. Not as promises — as mechanisms you can check.
SSH access, logged three ways
Command-line access runs over SSH with per-client keys. Every session is written to an operations log that exists in three places: on your machine, on ours, and in a version-controlled repository with timestamped history. What we did and when is verifiable — not just claimed.
Remote desktop, recorded end-to-end
When we operate your screen via RustDesk, the full session is recorded. The recording stays on your machine, so you can review exactly what was seen and done.
You can cut us off at any time
Revoke a single line in your SSH key file, or switch off Remote Login, and our access ends immediately. The logs remain — permanently, on your side — even after access is revoked.
A dedicated AI workstation — never your personal laptop
- We don't log into your personal laptop. Personal machines hold private material — banking, family, files unrelated to work. That boundary is not negotiable on our side.
- A dedicated machine does the work. Our standard setup: you purchase a new Mac mini or iMac as a dedicated AI workstation. It joins your network, runs the automation, and nothing else touches it.
- You own the machine. You buy it, you keep the receipt, you keep the warranty. We deploy remotely over a private tunnel (Tailscale) and set it up — the hardware stays yours, permanently.
Contracts, insurance & daily reports
Professional insurance
We hold Professional Indemnity (A$2M per claim) and Public Liability (A$20M) cover, underwritten through Lloyd's of London. A certificate of currency is available on request.
Sample contract on request
Every engagement runs on a written agreement — scope, milestones, payment schedule, data handling. Want to review the terms before talking? Ask and we'll send a sample contract over.
A daily report whenever we log in
If we logged into your systems — whether we changed something or only inspected it — you receive an email that same day stating what was done, what was found, and how anything we touched can be rolled back. No silent access. Ever.
AI-specific boundaries
- Client documents are processed within workflows we build for the client — not pasted into third-party AI tools for our own convenience
- Where AI services are used in a workflow, the data flow (what goes where, retained how long) is documented in the project proposal
- Sensitive categories — patient clinical data, legal privilege, regulated financial data — are identified during discovery and receive a dedicated data-risk review
- We will decline workflows that require unauthorised handling of sensitive personal data, even if technically straightforward
Retention & your rights
Form submissions are retained only as long as needed to handle your enquiry. You can request access to, correction of, or deletion of your personal information at any time by emailing [email protected]. See our Privacy Policy for the full statement.
Last reviewed: October 2026 · Maintained by Snail AI · This page is reviewed whenever our practices change — not on a schedule for its own sake.