Only what helps us serve you.
When you become a patron, we collect the information needed to do the work — your name, your email, your business details, the patrons you serve, and the work product Bob and Alice generate on your behalf.
When you visit the workshop without becoming a patron, we collect basic analytics — pages visited, browser type, anonymized IP — to understand how visitors find us.
We also record anonymous first-party usage events on this site, such as which pages and sections are viewed, which buttons are clicked and whether the contact form was sent. We set no cookies and record no personal data: never your name, email or what you write. Visits are counted with a hash that rotates every day, so we can’t follow you from one day to the next. If your browser sends Global Privacy Control or Do Not Track, we record nothing.
Data categories we may collect include contact details, account and billing records, business and project context you provide, work product generated during service delivery, support correspondence, and limited technical usage data (such as browser type, pages viewed, and IP-derived region). We use only essential site technologies and basic measurement tools needed to operate and improve the site.
When you connect Atelier to ChatGPT, Codex, Claude, or another MCP client, we also process the account identity needed to authenticate you, the site files you ask Bob to read or edit, audit records of tool calls, and the live site URL and repository slug associated with your account. We use that data only to confirm that you own the site, perform the requested edit, keep an audit trail, and recover or publish the site when asked.
To do the work.
The information we collect is used for one thing: to deliver the patronage you’ve hired us for. Bob needs your business context to build your site. Alice needs your patron list to draft follow-ups. The principal needs your history to give you good counsel.
We don’t repurpose your data. We don’t enrich it for advertising. We don’t sell it. We don’t share it with anyone other than the service providers we need to operate (hosting, payments, email delivery, authentication, database storage, source control, and secure traffic routing), who are bound by confidentiality.
For connected-app site editing, those providers may include Clerk for login, Supabase for account/site mapping, GitHub for the customer site repository, Cloudflare for secure routing, and OpenAI or Anthropic as the client surface you choose to use.
Our processing is based on service delivery, legitimate operational interests, and legal obligations where required. If a jurisdiction requires consent for specific processing, we rely on consent for that processing.
Three commitments.
- We don’t sell your data. Not to anyone, ever.
- We don’t track you across the internet. No third-party advertising trackers, no fingerprinting, no cross-site identity stitching.
- We don’t enrich your data with third-party sources. What you tell us is what we know.
We’re a processor, not a controller.
When Alice manages your CRM or your customer comms, the data she handles is your patrons’ data — and you’re the controller of it. We process it on your instruction and don’t use it for any other purpose.
Where required, we handle that data under a written data-processing agreement describing confidentiality, security, subprocessors, and deletion/return obligations.
Otto Cloud.
When you subscribe to Otto Cloud at cloud.elier.ai, we process:
- Your account. Your name and email, through our sign-in provider (Clerk), and the names and emails of people you invite. Invitation emails are sent through Resend.
- Billing. Your subscription and payment status. Payments are handled by Stripe; we never see or store your full card number.
- Your agent’s data. Your conversations, your agent’s memory, files and settings, and the credentials for any model accounts you connect, which are stored encrypted and used only by your agent.
Each customer’s agent runs in its own isolated container with its own storage on Cloudflare. One customer’s agent cannot read another’s data. Your agent’s data is backed up automatically to Cloudflare storage while it’s in use, so it can be restored if something fails.
To answer you, your agent sends your messages and the context it needs to the model provider in use: Anthropic or OpenAI when you connect your own account, or OpenRouter and the model provider it routes to for the included credit. Those providers process prompts under their own terms and privacy policies.
We don’t sell your agent’s data. When you cancel and ask us to delete your agent, we delete its data and backups.
Questions about Otto Cloud and your data: hello@elier.ai.
How to reach us about your data.
To request a copy of your data, ask us to delete it, or correct it, write to us:
We respond within seven days. If you’re in a jurisdiction with formal data-subject rights (GDPR, CCPA, etc.), we’ll handle requests within the timeframes those laws require.
You may request access, correction, deletion, export, or restriction of your personal data, and may object to certain processing where local law provides that right. If local law gives you an appeal path or regulator complaint right, you may use it.
Updates to this notice.
When we make material changes to how we handle data, we’ll update this page and tell patrons by email. The effective date below tracks the most recent version.
Effective date: 2026-05-31. Otto Cloud section added 2026-10-27.
hello@elier.ai · privacy@elier.ai