Last updated: July 25, 2026
Tenfold asks for a lot of trust. You are connecting an AI employee to the tools you run your business on, and then letting it do work without watching every step. This page explains exactly what that means, what protects you, and where we are still building.
Everything you send us travels over TLS and is encrypted on disk with AES-256.
Each account gets a dedicated, network-isolated namespace. Your AI employees never share a runtime with another customer.
Your work is not used to train our models or anyone else’s, and we never sell it.
Connecting a tool grants only the permissions that tool needs. Disconnect it and the access is gone immediately.
Employees can be required to check with you before they run anything, and you choose where those requests reach you.
Removing an employee tears down its environment and revokes its model key. Closing your account removes your data.
Where your data lives
Tenfold runs on Amazon Web Services in the United States. Your account data, the work your employees produce, and the files you upload are stored in a Postgres database and object storage managed by Supabase, also in the United States.
All traffic between you and Tenfold is encrypted in transit using TLS. Data at rest is encrypted with AES-256. Access to production systems is restricted to the people who need it, and every connection to your data from the application is enforced by row-level security in the database, so one account cannot read another's rows even if application code is wrong.
Your AI employees run in their own space
When you create a workspace, we provision a dedicated, network-isolated Kubernetes namespace for it. Your employees, the files they work with, and the credentials for the tools you connect all live inside that namespace. They are not shared with any other customer, and they are not pooled into a common agent runtime.
Each account also gets its own AI model key with its own spending budget. That key cannot be used by another account, and it is the mechanism that stops runaway spend as well as cross-account access.
What we do, and do not do, with your data
- We do not train on your data. Nothing you or your employees write is used to train our models.
- We do not sell your data or share it with advertisers.
- Your work is yours. Everything your employees produce belongs to you, including if you cancel.
- We route model calls through our own proxy so that usage is metered against your credits and logged for your review, rather than sent directly from an employee to a model provider.
Connecting your tools
Integrations use OAuth wherever the provider supports it, so Tenfold never sees your password. We request the narrowest set of permissions that lets the work happen, and we tell you what those permissions are before you approve them.
Access tokens are stored encrypted and scoped to your workspace. You can disconnect any integration from your settings at any time, which revokes Tenfold's access immediately. You can also revoke access from the provider's own security settings, and we will stop working rather than try to reconnect.
What an AI employee can and cannot do
An employee can only reach the tools you have connected and the knowledge you have given it. It cannot browse your other accounts, and it cannot grant itself new permissions.
For actions that run commands rather than just read and write through an integration, Tenfold supports approvals: the employee stops and asks before it proceeds. You control whether approvals are required and where the request reaches you, including as a direct message in Slack. Every job an employee runs is recorded, so you can see what was done, when, and what it produced.
We will always tell you plainly what is automated and what is not. If an employee does something wrong, contact us and we will help you understand what happened from the job history.
Retention and deletion
You can delete individual work, remove an employee, or close your account. Removing an employee tears down its environment and revokes its model key. Closing your account removes your workspace data from our production systems, and it ages out of encrypted backups on our normal backup rotation.
We keep the minimum records we are required to keep for billing and tax purposes.
Who else touches your data
We use the following sub-processors to run Tenfold. Each one receives only what it needs to do its job.
| Sub-processor | Purpose | Location |
|---|---|---|
| Amazon Web Services | Application hosting and compute | United States |
| Supabase | Database, authentication, and file storage | United States |
| Anthropic, OpenAI, Google | AI models that perform your work | United States |
| OpenRouter | Model routing | United States |
| Stripe | Payments and subscription billing | United States |
| Slack | Messaging, where you connect it | United States |
| Mailgun | Transactional email | United States |
| PostHog | Product analytics | United States |
| Datadog | Application monitoring | United States |
Where we are still building
Tenfold is an early-stage product and we would rather tell you that than imply otherwise.
- We do not hold a SOC 2 report today. If you need one before you can buy, tell us, because that demand is what decides when we start the process.
- We do not yet offer single sign-on, audit-log export, or a signed data processing addendum as self-serve features. Get in touch if you need any of them and we will work out what is possible.
- We do not currently offer data residency outside the United States.
Reporting a vulnerability
If you believe you have found a security issue, email security@hiretenfold.com. Tell us what you found and how to reproduce it. We will acknowledge your report, keep you updated while we fix it, and we will not pursue action against anyone who reports a problem in good faith and gives us a reasonable chance to address it.