-
Tokens and the prompt cache, explained
A new page explains how each provider's prompt cache works, what Codelux does when your AI comes back to a project, and where the app shows the tokens and the cache of your own sessions.
-
Codelux 4.8: memory as proof
The agent's memory now carries the code it refers to. When work resumes, Codelux checks that code against the real files, so the agent can trust what it reads, and keeps it up to date when another agent edits it.