Lexxy, a personal AI companion that remembers: journaling, patterns and meeting memory
Lexxy is a daily AI companion that learns the user's habits, lifestyle and language over time. It journals with them, surfaces patterns they would miss, and powers a Soundroom that records meetings and reflects back on tone, decisions and credibility. Every conversation becomes a memory the user owns.
Daily
Journaling that builds on yesterday
Soundroom
Meeting memory with reflection
Owned
Memory belongs to the user
The constraint
A companion is only useful if it remembers. Most chatbots reset.
Lexxy set out to build a daily companion rather than an assistant you prompt. The product promise was that the AI would know the user better each week: their habits, their language, what they were working through. A conversation that forgot everything at the end of the session could not keep that promise.
The second half of the brief was meetings. Users wanted a record of what was said and decided, and an honest reflection on how they came across. That meant speech capture, a memory that spanned journal entries and meetings, and a product where the user, not the vendor, owned the memory.
What we built
Memory first, then the journal and the Soundroom on top of it.
We designed the memory architecture before the features. Every journal entry, conversation and meeting is written to a memory the user owns and can read, so the companion's picture of them is inspectable rather than hidden. Daily journaling builds on what came before, and a pattern layer surfaces what the user would miss on their own: a recurring worry, a habit that shifted, a decision they keep deferring.
The Soundroom records meetings through a speech pipeline and reflects back on tone, the decisions taken and the user's own credibility in the room. Those reflections join the same memory, so a journaling session the next morning already knows how the meeting went.
- User-owned long-term memory across journal, chat and meetings
- Daily journaling that references and builds on history
- Pattern surfacing over weeks and months
- Soundroom meeting capture with tone and decision reflection
- An AI that grows with the user rather than resetting
What changed
A companion that knows the user better each week, and a memory they can take with them.
Lexxy ships as a daily companion with journaling, pattern surfacing and meeting memory in one product. The memory layer is the product's moat: it is what a user would lose by leaving, and it belongs to them.
What the team owns
What was running, and in whose hands, when we stepped back.
- The memory architecture, the journaling and pattern features, and the Soundroom pipeline
- Model integrations with the prompts and evaluation sets behind them
- Documentation for the product team to extend features without us
Product build for Lexxy