Security
Because there aren't any. untactit authenticates through your existing identity provider, which means there is no password database here to breach.
Sign-in goes through Google, Microsoft, or your SAML provider. We never see, hold, or transmit a password — there is no credential table to steal.
Every record carries its organization. Queries are scoped at the data layer, not filtered in the interface — a missing filter fails closed, not open.
Who changed what, when, and from where. Entries are written once and never edited — including by us.
Permission is decided by the server on every request. The interface reflects that decision — it never makes its own.
Identity providers
TLS on every connection reaching the service from outside.
Vendor credentials and other sensitive fields are encrypted at the application layer with a separate key, so a database dump alone does not expose them.
Encryption keys are versioned. A new key can be promoted while previous versions stay readable, so rotation does not require downtime or a migration window.
None stored for SSO users. Where a local credential exists it is a PBKDF2-HMAC-SHA256 hash with per-record salt, never reversible.
Disk-level encryption stops a stolen drive. It does not stop a leaked backup or an over-broad query, so we encrypt sensitive fields separately.
An audit log is only worth something if you can prove nobody edited it. Ours chains each entry to the hash of the one before it — change a single row and every row after it stops verifying.
What the log answers
Changes are implemented against a development bundle that is never reachable from the public internet.
Syntax and symbols, type checking, real-data verification, runtime behaviour, a security audit, dependency vulnerabilities, and a scan for credentials that ended up somewhere they can leave from. A partial pass does not promote.
One command moves the verified bundle to production, and one command puts the previous release back.
Review is not left to whoever is looking. Fixed criteria hold the veto, they run on every change, and they do not get tired.
Most of what runs in any product is somebody else's code. We keep a bill of materials for ours and check it against a public vulnerability database.
Found something? Write to our security address. We will confirm receipt and tell you what we are doing about it.
Who else touches your data
Tools you connect run on your own credentials and your own agreement with that vendor — they are not our subprocessors. We give 30 days' notice before adding one of ours.
Security pages usually list certifications. Here's where we actually are.
Under NDA we share the detail this page does not carry: a full CAIQ v4 self-assessment across all 261 questions, our architecture and data-flow documentation, subprocessor list, and the DPA. Security reviewers get the complete picture — it just isn't published to the open internet, which is the same standard we hold your data to.
Security questions, disclosure reports, or a DPA request — get in touch. We answer security mail before we answer sales mail.
Connect one workspace and see every skill, rule, and memory your team has in play — in about ten minutes.
No credit card. Works with what you already run.