Download Lutin

v0.49.0 makes Lutin better at carrying conversations forward, asking memory follow-up questions more naturally, and keeping those answers tied to the right context across desktop and mobile. It also fixes a note-creation issue that could create duplicates and raises the daily allowance for free users.

Highlights

• Lutin can now continue the most recent conversation from the main composer, while still separating a message into a new thread when it is clearly a new topic.
• Follow-up replies are routed more reliably: Lutin can keep the current thread, join a different recent thread, or start fresh before answering.
• Memory clarification questions now appear as natural Lutin-authored messages instead of raw system prompts.
• Memory clarification answers work more smoothly on desktop and mobile, including quick replies like “I own it,” “I work on it,” or “Just research.”
• Free plan daily usage increased from 100,000 to 1,000,000 tokens.

Fixes and Improvements

• Fixed a bug where creating a new note from chat could create duplicate notes when the note panel remounted.
• Memory clarification buttons now retire when the question expires or when the answer is resolved elsewhere.
• Answering a memory clarification now sends a normal user reply in the conversation, so the resolution is easier to follow in the feed.
• Lutin no longer guesses which memory question a short free-form answer belongs to unless that answer is connected to the relevant clarification.
• Mobile now supports the same memory clarification answer flow as desktop.
• Mobile conversation replies can now automatically include the previous conversation context when no explicit feed item is attached.
• Feed handling for memory clarification events is more consistent across desktop and mobile.
• Added stricter frontend checks around desktop entry points to catch UI lifecycle issues earlier.

Internal

• Added conversation-thread tooling so the assistant can inspect recent threads, read thread history, and attach or detach the current request before responding.
• Improved event-thread persistence for conversation continuation, including preserving user request content when moving a request between threads.
• Memory clarification events are now processed as assistant-authored outbound interactions, improving supportability and avoiding raw internal prompts in the user experience.