Safety Lab Aero is built so the AI accelerates the work without ever becoming part of the safety argument. The deterministic engine owns every safety‑significant number, the AI is advisory and gated, and controlled documents are routed only where you allow them.
Failure rates, probabilities, cut‑sets, and DAL allocation are computed by explicit, repeatable logic — never by the AI. The AI may draft structure, but it never sets λ, DAL, or severity as fact.
Every AI edit is proposed, not applied: you see a before‑to‑after diff, accept what you want, and undo any turn. Nothing the AI does is silent or irreversible.
The project model and the deterministic engine run in your browser or desktop app; your projects are your own files. Only an explicit AI request sends anything to your configured AI backend.
You choose where AI requests are processed, to match your programme's data‑handling posture. Documents you mark controlled (ITAR or proprietary) are hard‑routed: while one is in the project, every cloud AI path is blocked — server‑side — unless the backend is on‑premise.
| Backend | What it is | Controlled docs |
|---|---|---|
| Hosted proxy | Claude via the Pro+ hosted proxy. | Not permitted |
| Bring‑your‑own key | Your provider key, used directly. | Not permitted |
| ITAR (US‑sovereign cloud) | Planned: hosted proxy routed to Azure Government. Until live, the proxy refuses ITAR traffic server‑side rather than routing it to any public‑cloud model. | Planned |
| On‑premise | Self‑hosted / local backend. | Permitted |
Controlled‑document routing is enforced in the application: a controlled document on file blocks every AI call — chat and batch alike — unless an on‑premise backend is configured (US‑sovereign cloud routing is planned), and where each controlled document was processed is recorded. The block is enforced server‑side at the proxy, not only in the browser. This is a routing control, not a claim of formal accreditation; talk to us about your specific ITAR/EAR or programme‑security requirements.
No. Those are owned by the deterministic engine and the engineer. The AI may propose a severity with rationale for you to confirm, and may build fault‑tree structure with rates left blank — but it never establishes a safety number as fact.
No — by design. The AI drafts, traces, and connects; the deterministic engine and the engineer own the numbers, the allocation, and the safety case. Requirements are allocated by interpreting the fault‑tree logic, not by the AI.
The model and engine run locally in your browser or desktop app. Only an explicit AI request sends the relevant context to your configured backend — hosted proxy, your own key, or on‑premise. For ITAR projects, cloud backends are blocked server‑side (US‑sovereign cloud routing is planned).
Mark the relevant documents controlled; the AI is then restricted to an on‑premise backend and blocked from every cloud path, enforced server‑side (a US‑sovereign cloud option is planned). For fully air‑gapped work, the on‑premise backend keeps everything inside your boundary.
No. It is a tool your engineers and DER use; the human owns every judgement and signs the safety case. The platform's job is to make the work faster and the golden thread complete and traceable.
ARP 4754B and ARP 4761A, the means of compliance for your cert basis (AC 23.1309 / AC 25.1309 / AC 27 / AC 29 / SC‑VTOL, ASTM F3230), and DO‑178C / DO‑254 development assurance.
Advisory AI above a deterministic, engineer‑owned core — with the data‑handling controls a controlled programme needs. Questions about your specific security posture? We're happy to talk.
Open the app →