On this page
A plain-language summary of how we protect what you put in your Suite. The binding commitments are in Section 10 of the Subscription Terms and Annex A of the Data Processing Addendum.
Where your content lives
- Your own private repository. Each Suite's content is stored in its own private GitHub repository, separate from every other customer's.
- Your own application. Each Suite runs as its own application, with its own database, on Sevalla (a Kinsta platform) on Google Cloud in the us-west3 region (Salt Lake City, Utah).
- Stored in the United States. We store and process your content in the United States. Two exceptions: Cloudflare, which carries traffic to the Suites, may route it through its global network, and the AI processing Anthropic does for us can run in other countries. The full list is at storystak.com/legal/subprocessors/.
Who can see it
- Sign-in with Google. Your team signs in with Google. On every request, the Suite re-checks the signed-in email address against your Suite's list of allowed domains and people, so deactivating someone in your Suite ends their access. Shared sign-ins are not allowed.
- Client walls. Each client's Stak is kept separate from your other clients'. You can limit a team member to the clients you assign them, and they then cannot read any other client except one they manage, from their AI tool or in the Surface. Each read through an AI tool covers one Stak.
- Switched-off clients. A client you switch off can't be read by your team's AI. Your owners and admins can still open it.
- Limited access at Storystak. Only people at Storystak who need access to run, secure or support the Service have it. Today that is Storystak's two partners; contractors get access only for a named customer task.
How content changes
- People approve changes. AI tools can read your Stak and propose changes, but no change applies until a person with that right approves it, and each change records who proposed it.
- Some changes always need a person. Any number, price, percentage or date, and the files that speak for a client, can never be set to apply on their own.
- Secrets are turned away. A proposed change that contains what looks like an access key or a payment-card number is refused when it is filed.
In transit
Your Suite, the connector and Dot are served only over HTTPS with TLS 1.3. Plain HTTP is redirected to HTTPS, and browsers are told to use HTTPS only.
What we don't do with it
- We don't sell your content.
- We don't use it to train AI models, and Anthropic is our AI provider. Anthropic works under terms that don't allow it to train on your content. Dot's voice provider is set not to use Dot calls to train its models.
- We don't use one customer's content to build or improve anything for another.
Dot
Dot's voice provider transcribes calls and is set not to keep the audio, and we don't keep it either. We and the voice provider each keep a call's transcript for no more than 30 days.
Deletion
When you ask, we delete your content from the live Service within 30 days. Within 90 days after that, backups containing it expire and we take reasonable steps to remove it from version history.
If something goes wrong
We will tell you within 72 hours of confirming that someone got unauthorized access to your content. To report a security problem, email security@storystak.com.
What we don't claim
- No certification for Storystak. Storystak's own system has not been independently audited or certified. Our hosting provider, Sevalla, holds a SOC 2 Type II report for its own infrastructure, and we can help you request it under Sevalla's confidentiality terms. That report covers Sevalla's controls, not Storystak's application.
- Reads aren't logged one by one. We record which tools each person uses and on which days, and every change to your Stak is recorded. We don't keep a log of each individual read.