Glossary
What is stateless AI?
A stateless AI keeps no memory between requests: each message is answered from what is in front of it and nothing is carried forward, so there is no accumulating record of you anywhere in the system.
Stateless does not mean forgetful mid-conversation
The distinction confuses people. Within a conversation the model still sees the earlier messages — they are re-sent with each turn, held by your browser rather than by a server.
What is absent is memory across sessions. Close the tab and the context is genuinely gone, because the only copy was in front of you.
What it costs and what it buys
The cost is real: no assistant that remembers your projects, your preferences or last week’s conversation. Products built on cross-session memory are more useful for exactly that reason, and if that is what you want this architecture is the wrong one.
What it buys is that there is nothing to breach, subpoena, sell or train on — not as a policy, but as a property of the design. It is the difference between promising not to read your file and not having one.
Asked before you asked
Does stateless mean the AI forgets mid-conversation?
No. The conversation so far is re-sent with each message, so it follows the thread. It just never stores that thread anywhere after you close the tab.
Why do most AI products keep state?
Because memory makes an assistant more useful, and because it is easier to build. Statelessness is a deliberate constraint accepted for what it removes.
Can a stateless service still log requests?
Yes — stateless describes the conversation model, not the logging policy. Both need to be true, which is why the retention question is asked separately.