Security and Trust
Last updated October 24, 2026
The plain answers to what security and IT teams ask before they connect their business to Wrenhouse: where it runs, who processes your data, which permissions we ask for and why, and what happens if something goes wrong. Our Privacy Policy and Terms of Service have the full detail. If you need a security questionnaire filled in or a data processing agreement, email support@wrenhouse.io.
1. At a glance
- Hosted in the United States, on Render, which runs on Amazon Web Services in Ohio.
- Encrypted in transit everywhere. The keys to your connected tools are encrypted again inside our database.
- Every hub is kept separate. No customer can see another customer's data.
- Your content is never used to train AI, ours or anyone else's.
- Staff open a hub only to help, tied to a ticket or reason, and every visit is logged.
- Agents draft. Nothing is sent, posted or changed in your tools without a person's approval, unless your super admin turns on automatic mode for that job.
- We tell you without undue delay if your data is ever exposed, and within 72 hours of confirming it whenever we can.
2. Where Wrenhouse runs
The app at app.wrenhouse.io and its database are hosted by Render in its Ohio region (Amazon Web Services us-east-2), in the United States. Render runs the servers, network and managed PostgreSQL database. All your hub's data is stored there.
Your hub's files and records are stored only in that region. The AI provider receives just the text each task needs and returns the result.
3. Who processes your data
These providers process data for us. Each one gets only what it needs to do its part.
| Provider | What it does | What it sees | Where |
|---|---|---|---|
| Render | Hosts the app and database | All hub data, stored encrypted at rest | United States (AWS, Ohio) |
| Anthropic | Claude AI models that read, sort and draft | The text needed for each task, not kept for training | United States |
| Our own email (Google Workspace), and Google sign-in when you choose it | Emails you send us, and your sign-in details | United States and worldwide | |
| Mercury | Business banking, invoices and payments | Your billing contact and invoices | United States |
| Resend | Account emails, such as invitations and password resets, once switched on | The recipient's email address and the message | United States |
| Cloudflare | Relays live help voice calls when a direct connection isn't possible | Encrypted call audio it can't decode, and network addresses | United States and worldwide |
Before a new provider handles your hub content, we add it here and to the Privacy Policy and tell your super admin.
4. Permissions we ask for
When you sign in with Google or Microsoft, or connect a tool, that provider shows you a consent screen listing what Wrenhouse is asking for. Here is every permission and why. Learning from a mailbox or calendar is a separate switch inside the hub, off until you turn it on, and learning only ever reads.
Signing in with Google or Microsoft
| Permission | Why |
|---|---|
| Your name, email address and profile picture (openid, email, profile) | To sign you in and show who you are. Signing in gives us no access to your email, files or calendar. |
Connecting Gmail
| Permission | Why |
|---|---|
| Read your email (gmail.readonly) | So agents can sort your inbox and draft replies |
| Manage drafts and send email (gmail.compose) | To save the drafted reply in your Gmail, and to send it only when you press Send |
| See your contacts, and people you've emailed (contacts.readonly, contacts.other.readonly), optional | To suggest the right people in To and Cc |
Connecting Google Calendar
| Permission | Why |
|---|---|
| View and edit events (calendar.events) | To show your day in the hub and, only when you approve, add things like a review block or a booking |
Connecting Microsoft 365 or Outlook
| Permission | Why |
|---|---|
| Sign you in and read your profile (openid, email, profile, User.Read) | To know which account is connected |
| Stay connected (offline_access) | So the hub keeps working without asking you to sign in every hour |
| Read and write mail (Mail.ReadWrite) | To read your inbox and save drafted replies in Outlook. Microsoft has no drafts-only permission |
| Send mail (Mail.Send) | To send a reply only when you press Send |
| Read and write calendars (Calendars.ReadWrite) | To show your day and, only when you approve, add events |
| Read contacts and relevant people (Contacts.Read, People.Read) | To suggest the right people in To and Cc |
Connecting a CRM
| Tool | Permission | Why |
|---|---|---|
| HubSpot | Read contacts, companies and deals | So agents know your accounts and pipeline. Read only |
| Salesforce | API access and stay connected (api, refresh_token, openid, id) | Salesforce offers no narrower read-only option for apps. The hub only reads, and your Salesforce admin approves the app first |
Other email (IMAP)
For any other email provider, you give the hub your address and an app password, never your main password. It reads the inbox, keeps drafts, and sends only when you press Send.
Each person controls their own mailbox and calendar connections, and a mailbox's messages are shown only to the person who connected it. More tools, such as help desks and issue trackers, will be listed here before they go live.
Google user data. Wrenhouse's use and transfer of information received from Google APIs to any other app will adhere to the Google API Services User Data Policy, including the Limited Use requirements.
5. How to remove our access
- In the hub: open your connections in Settings and choose Disconnect. We delete the stored access straight away and, where the tool allows it, revoke it with that tool too.
- Google: open your Google Account, go to Security, then "Your connections to third-party apps and services", choose Wrenhouse and remove access.
- Microsoft: for a work or school account, open My Apps or ask your Microsoft 365 admin to remove Wrenhouse. For a personal account, open your Microsoft account's privacy settings, then "Apps and services".
- HubSpot or Salesforce: remove Wrenhouse from the connected apps in that tool's settings.
- Learning: turn it off for any source on the "What your hub knows" page, and delete what was learned.
6. How your data is protected
- In transit: every connection to Wrenhouse and from Wrenhouse to your tools uses HTTPS (TLS).
- At rest: the database is encrypted at rest by our host. On top of that, the keys and tokens for your connected tools are encrypted with AES-256-GCM using a secret held only in our server settings, so a copy of the database alone can't unlock them.
- Passwords: stored only as a salted hash. Nobody at Wrenhouse can see your password, and we never ask for it in chat or tickets.
- Sessions: sign-in cookies are secure and HTTP-only, and last up to 30 days or until you sign out. Resetting your password signs you out everywhere.
- Sensitive details: card numbers, bank details, passwords and ID numbers are masked before the hub learns anything from your content.
- Least access: your super admin decides who joins your hub and what each person can see. Wrenhouse staff permissions are limited by role.
- Limits: sign-in and other sensitive actions are rate limited to slow down guessing and abuse.
7. Keeping every hub separate
Every customer's hub is its own workspace. Every request is checked against the signed-in person's hub, and every lookup in the database is limited to that one hub, so no customer, person or agent can reach another customer's data. What the hub learns about your business stays in your hub and is never used for anyone else.
8. When our staff open your hub
Wrenhouse staff open a hub only to help, for example on a support ticket, during onboarding or to fix a problem. Each visit is tied to a ticket or a stated reason and recorded with who, when and why. Your super admin can see this access log, and it is being added to your hub's Settings so you can check it any time. During a live support session a banner shows, and you can see where we are pointing. Taking control of your screen needs a separate yes from you, and you can stop it at any time.
9. AI and your data
- Wrenhouse uses Anthropic's Claude models through Anthropic's business API.
- Under Anthropic's commercial terms, what we send isn't used to train its models. We don't train any models on your content either.
- Only the text needed for a task is sent, and the result comes back to your hub.
- Agents draft and a person approves by default, so a person checks AI work before it goes out.
10. If something goes wrong
We watch for errors and unusual activity and fix problems quickly. If your data is ever exposed, we tell your organization without undue delay, and within 72 hours of confirming it whenever we can: what happened, what it affects and what we're doing about it. We also notify regulators and the people affected where the law requires.
11. Backups, export and deletion
- Backups: the database is backed up by our host with point-in-time recovery, so we can restore it after a failure.
- Export: ask for a copy of your hub's data at any time, and for 30 days after your agreement ends.
- Deletion: after those 30 days, your hub data is deleted. Copies in backups age out within 30 more days.
- Disconnecting a tool deletes its stored access immediately.
12. Certifications
Wrenhouse is a young company and doesn't hold a SOC 2 report or ISO 27001 certification yet. Our host, Render, and our AI provider, Anthropic, each hold their own independent security certifications, including SOC 2 Type II. We'll list our own here when we have them. Until then, we're happy to answer your security questionnaire in writing.
13. Report a security issue
If you think you've found a security problem, email support@wrenhouse.io with "Security" in the subject. Please don't access other people's data or disrupt the service while testing. We'll reply quickly and keep you updated until it's fixed.
This page describes how we protect information. It isn't legal advice, and it doesn't change our Terms of Service.