Security & data handling
Cryogram is built so that the sensitive path does not exist, rather than promising not to use it. Your automations are built and run on your own machine, on your own AI key, and the parts of the product that involve our infrastructure are strictly limited to what you choose to publish. This page explains how that works, in plain language, for the two halves of the product: the Cryogram Builder, which runs on your machine, and Cryogram Cloud, the optional hosted service.
Cryogram Builder
Your data stays in your systems
Automations are stateless: each run reads inputs, writes outputs, and keeps none of your business data in between. The documents, records and messages an automation touches live in your own files and systems, and Cryogram does not keep a copy. What Cryogram stores on your machine is its own operating state: the automation definitions, run history and settings.
Build material never leaves your machine
The material generated while building - your chat with the builder, sample files, and the recorded evidence each automation is tested against - is stored locally and is never uploaded, not even if you later publish the automation to Cryogram Cloud. What you send to the AI during a build session goes directly to your AI provider on your own key, under your own agreement with them.
Secrets: the AI sees names, never values
API keys, passwords and tokens are stored locally in a restricted file with owner-only permissions, are never displayed back in the interface after entry, and are never included in anything an AI model reads. Automation steps receive secrets by name; the values are injected at the moment a step runs and are scrubbed from outputs and logs. There is no code path that passes a secret value to a model.
Approval gates on anything external
Steps that act on the outside world - sending email, writing to a CRM, posting to an API - sit behind an approval gate. For each automation, you decide whether such steps need a human confirmation on every run or are approved to run unattended. Irreversible actions are flagged as such.
Every step checks its work
No value moves from one step to the next without passing that step's checks, which are derived from proven runs. When a check fails, the run stops cleanly and can resume from where it paused once the problem is resolved. An automation that cannot run correctly stops and tells you; it does not improvise.
Repairs are tested before they ship
When something changes and a step breaks, the AI investigates and proposes a fix. The fix is promoted only after passing the automation's tests against the recorded evidence of past good runs. A repair that cannot prove itself does not go live.
Cryogram Cloud
What moves to the cloud, and what never does
Publishing an automation moves exactly what the published version needs to run: the automation itself, its schedule and trigger configuration, the named secrets it uses, and any static data files - all shown to you in a plain summary before you confirm. Your builder chat, samples, recorded evidence and your own AI key never leave your machine; the cloud has no code path that could use a builder key, and never holds one.
Isolation
Each published automation runs in its own isolated environment, with hardware-level isolation between customers, resource limits, and no inbound access except through our gateway. Outbound network access is denied by default: each automation declares the domains it needs, and only those are reachable.
Encryption and secrets in the cloud
Data is encrypted in transit and at rest. Secrets you consent to sync are encrypted per customer and injected into the automation's environment only at start - the same names-not-values contract as on your machine. Secrets never appear in logs, run records or stored files.
Data residency
Customer data at rest - published automations, run results and stored files - can be hosted in your choice of region, including the EU where that is required. A data processing agreement and the current subprocessor list are available to Cryogram Cloud customers.
Retention and deletion
Run results are retained for 90 days and system logs for 30 days, both documented and subject to change with notice. Unpublishing an automation tears down its environment and deletes its cloud data, stored results and logs after a short grace window.
Auditability
Every run produces an append-only event log: what ran, what each step received and produced with secrets scrubbed, what stopped and why, and who approved what. Published versions are immutable - only re-publishing changes what runs - and previous versions are kept so a change can be rolled back.
Your AI provider
Cryogram is not a model provider. During building and repair, content you choose to work with is processed by your AI provider (for example Anthropic) on your key, under your agreement with them. Once an automation is built, only the steps that genuinely need judgement call a model at all; the routine majority of a run calls no model.
This website
The waitlist form is operated by Tally, our form processor; what we collect and why is described in the privacy policy. This site sets no advertising trackers.
Reporting a vulnerability
If you believe you have found a security issue in Cryogram or this site, email enquiries@scopecreeper.com with “SECURITY” in the subject line. We read every report and will respond promptly. Please give us reasonable time to remediate before public disclosure.