Intelligent prompt-based memory extraction

That phrase in the feature list is auto-capture: after a conversation, a background AI request summarizes technical work and saves it as memory. No special prompt from you is required. It uses opencodeProvider / opencodeModel when set, otherwise the manual memoryProvider fallback.

User profile

The User Profile is a separate, cross-project summary of how you like to work (preferences, habits). It is updated on an interval (userProfileAnalysisInterval, default every 10 analyzed prompts), shown in the web UI’s profile view, and readable via memory({ mode: "profile" }). You do not populate it by hand for normal use — profile learning fills it when a provider is ready. Output language follows autoCaptureLanguage (default "auto", mirroring the language of your prompts), the same setting used for auto-captured memories.

“No profile found. Keep chatting to build your profile.” is the expected empty state, not a crash. Profile learning needs:

1. Auto-capture running with a reachable provider (opencodeProvider + opencodeModel, or a complete manual fallback with memoryModel + memoryApiUrl).
2. Enough session prompts since the last analysis — at least userProfileAnalysisInterval (default 10).
3. That provider to support structured/tool-call output (same requirement as auto-capture).

If you have chatted for a while and still see the message, check that auto-capture is actually firing in the logs and that the configured provider succeeds (failed profile analysis no longer hides behind a generic empty state when the provider errors).