Every teammate runs a personal AI memory on their laptop. ForAll connects them — opt-in, local-first — so the smartest answer in your company is one keystroke away.
A passive desktop agent captures meetings, chat, docs. Agentic memory structures it. ⌘K answers in natural language — with sources, and who else knows more.
Meetings, chat, docs, browser, terminal. Passive desktop agent.
Structures the chaos into a knowledge graph. Knows what matters.
Natural language Q&A with citations and who knows more.
Opt-in across teammates. Personal memory becomes company brain.
Decisions, context, and ownership — scattered across Slack, meetings, docs, and people's heads. The smartest answers in your company are unreachable the moment they're spoken.
"What did we decide about pricing?" Re-litigated every quarter. The original thread is buried; the call wasn't recorded; the doc was never updated.
New hires repeat old mistakes. Why was the API designed this way? Why did we drop that vendor? The reasoning is in someone's head — or no one's.
"Who knows about X?" Tribal knowledge sits with whoever was there. When they leave, it leaves. When they're on PTO, the team stalls.
A passive local agent on each laptop. An agentic memory that decides what to keep. A ⌘K answer with sources. And — when teammates opt in — a network that compounds.
Meetings, chat, docs, browser, terminal. A passive desktop agent — nothing leaves your laptop.
The agent decides what matters, links it, and rebuilds the graph as your work evolves.
Ask anything in natural language. Get the answer with citations — and who knows more.
Opt-in across teammates. Personal memory becomes shared. This is where no incumbent can follow.
Each on its own is doable. The combination — local, personal, federated — has to be built from a different foundation.
Zero bytes leave your laptop unless you choose. The privacy answer is "it never went anywhere" — not "we encrypted it."
You get value before any teammate installs it. No empty-network problem. Personal memory is the product; federation is the multiplier.
When teammates install ForAll, their brains connect — opt-in per project. Suddenly you can query everyone's collective memory. This is the moat.
Ken built contextual AI memory at Meta. Mena ships products solo and runs production systems at scale. Between them, every primitive ForAll needs already exists in their hands.

Previously Meta Reality Labs Research, building contextual AI agents for always-on smart glasses. Designed "personal timelines" — compressing physical and digital context into queryable memory and predicting user intent. Deep experience in on-device ML, GPU infrastructure, and benchmarking vision-language models alongside human participants.

5+ years of enterprise mobile engineering at Canadian Tire Financial Services, leading scalable architecture, performance, and production reliability on platforms processing complex payments. Solo founder of multiple shipped products including a custom music streaming app and a fitness tracking platform — end-to-end product lifecycle, design through deployment.