Sub-processors
These are the companies that receive some part of your data so that Yeta Zeta can work. For each one: what it receives, and where it processes it.
This list is checked against the code. If a service is not on this list, the application cannot reach it.
Anthropic
Your request text and the work in progress, for the models that run on it. Token counts for pricing.
Processes in: United States
OpenAI
Your request text and the work in progress, for the models that run on it.
Processes in: United States
Supabase
Your email address and sign-in sessions. No run content.
Processes in: Mumbai, India (ap-south-1)
Neon
Everything else: your workspace, your runs, their answers and traces, and the usage ledger.
Processes in: Singapore (aws-ap-southeast-1)
Upstash
Rate-limit counters keyed on a salted hash of your IP address and on your user id. No IP address is stored.
Processes in: Mumbai, India (ap-south-1)
Cloudflare
A bot-check token from the signup form, and your IP address for that check only. We do not store it.
Processes in: Global edge network
Sentry
Error reports: the error type, the stack frame, and the runtime. Message text is dropped unless we have explicitly marked it safe.
Processes in: European Union (Germany)
Vercel
Requests to this site, including your IP address, as any web host does.
Processes in: Global edge network
Razorpay
Your payment details and billing contact. Card numbers reach Razorpay directly and never reach us.
Processes in: India
Where your run content goes
Your request and the work produced from it go to Anthropic and OpenAI, because that is what the product does — several models work on your request in turn. It is stored in our database at Neon in Singapore. It is not sent anywhere else.
What we do not say
We make no claim about whether these companies use your text to train their models. Their terms differ per model and change without notice, so any promise we made here could stop being true without us knowing. We would rather say nothing than say something we cannot keep.