Answered before you ask
Finish most of an assessment without a call — or rule us out in five minutes instead of week three. About a third of the answers are no.
Still deciding?Ask the people who build it.How is customer data isolated between tenants?
By Postgres row-level security, applied in the database engine against a session context set once per request — not by a WHERE clause each query is responsible for remembering. A query that omits the filter returns nothing rather than everything, and a caller who guesses another workspace's resource id receives a 404.Is data encrypted at rest and in transit?
Yes to both. TLS on every hop with HSTS on the public surfaces; storage-layer encryption on the database and object storage. Credentials for connected accounts get an additional layer: envelope encryption with a data key belonging to that workspace alone, plus a separate key for fields that must remain searchable.Do you train models on our data?
No. Content your agent reads and messages your visitors send are used to answer questions and to produce your own analytics. They are not used to train models, by us or by the model providers under the terms we hold with them, and they are not sold.Which subprocessors receive our data?
The full list is published and kept current, split into providers that are always in the path and those that only receive data if you connect them. It names what each one actually receives rather than a one-word purpose. See the subprocessors page.Do you offer a DPA?
Yes. We act as a processor for the conversation data your visitors generate; you remain the controller. Ask and we will sign one.What is your incident notification commitment?
A public write-up within 72 hours of confirming a customer-affecting security incident, plus direct email to every affected workspace — not just a status page update. The commitments are published in advance on the incidents page, including that entries are never quietly removed.Can we delete our data?
Yes. An account and everything in it, including conversations, contacts and files, can be deleted for good, from the primary database and from object storage.What happens to our data if we stop paying?
The agent stops answering, and your data remains available to export for a period before deletion. The specifics are in the billing policy.Do you have a vulnerability disclosure process?
Yes, with published commitments — including that we will not pursue legal action for good-faith research, will not require an NDA as a condition of fixing something, and will not silently patch and deny. Every report is acknowledged by a person within three working days.Is there an audit log?
Yes. Each workspace has an append-only audit log, and any access to your workspace by our staff is time-boxed, carries a stated reason and is written into that log. Conversations and every tool call, with its outcome, are recorded per workspace too.
Need your own form filled in? Send it to security@integrable.cloud · architecture · subprocessors · data residency