The notes app that does the filing
It is a Tuesday in 2026 and somewhere in San Francisco, a product manager is walking her dog. She talks into her phone for eleven minutes - half a sales meeting recap, two unrelated ideas, a grocery list that briefly turns philosophical. She doesn't tag anything. She doesn't pick a folder. She doesn't even pick a title. By the time the dog has eaten a discarded crust on Fulton Street, Mem has transcribed her ramble, split it into three notes, linked the sales recap to the customer's existing record, and placed the philosophical grocery list exactly where she will find it, which is to say, when she next searches for "the thing about almonds."
This is Mem in 2026: a small, opinionated company that decided your second brain shouldn't require a librarian. The product calls itself an "AI Thought Partner," which sounds like the kind of phrase a venture firm prints on a tote bag. The thing is: it largely earns the description. Mem doesn't just store what you write. It remembers it on your behalf, and resurfaces it the moment your future self stumbles back into the same problem.
Folders were a war crime
For two decades, productivity software made a quiet bargain with the user: we'll give you infinite storage, but you have to do the librarian work. Build a folder. Tag it. Cross-link it. Move it when the project name changes. The result was a generation of knowledge workers maintaining elaborate filing systems they didn't really trust and rarely revisited.
Kevin Moody and Dennis Xu were two of those workers. They were also, perhaps more painfully, two of the people who tried to fix it - again and again, with new apps, new tagging rules, new weekly reviews - and watched the system collapse every time. The frustration that became Mem was specific: information generated in one context was invisible in another. A note from a meeting on Monday simply did not exist when you needed it on Thursday. Worse, it might exist, somewhere, in a folder you no longer remembered naming.
The pitch deck version of this complaint is short. The lived version is exhausting.
Two people, one wager, several years of nerve
Moody and Xu started Mem Labs in 2019, before most people had typed a sentence into a large language model and watched it answer back. Their bet was, in retrospect, unreasonably early: if AI was about to make organization a computer problem, then the right thing to build was the user-facing tool that would benefit. Not a search box. Not a wiki. A notes product whose job was to make the act of organizing disappear entirely.
Andreessen Horowitz wrote a $5.6 million seed check in April 2021. A year and a half later, in November 2022 - mere days after ChatGPT's public debut would change the room temperature of every AI investor on Earth - the OpenAI Startup Fund led a $23.5 million Series A at a $110 million post-money valuation. Mem was one of the very first companies the fund had backed. The signal was loud. The expectation, naturally, became louder.
What you do (almost nothing). What it does (almost everything).
The current product, Mem 2.0, is a complete rebuild that landed after a long alpha. The shape is simple enough to fit on a napkin and stubborn enough to be hard to copy. There are five real surfaces.
Voice Mode
Talk into your phone. Mem transcribes, separates topics, and writes structured notes. No bot in your meeting required.
Email-to-Mem
Forward anything to save@mem.ai. It returns to your knowledge base as a clean, filed note.
Chrome Extension
Click once on any webpage. Mem captures, formats, and decides where it goes - so you don't have to.
Deep Search
Describe what you remember. "That conversation about Q3 budget concerns." It finds it. Keywords not required.
Heads Up
About to meet someone? Mem surfaces your last conversation with them before you walk in. Months of history, ten seconds of prep.
Mem API
For the developer who wants their own apps to feed Mem's brain. Pipe in, query back.
What ties these together is a refusal to make the user do the librarian's job. There is no "where should this go." There is no "did I tag this." There is a place to look, and a thing that resurfaces. Whether the model decides correctly every time is, of course, a different question - one Mem has been refining publicly for years, on the receiving end of both fans and skeptics.
A short history of one stubborn idea
Mem Labs founded in San Francisco
Kevin Moody and Dennis Xu start with a question: why is organizing information still the user's job?
Seed round: $5.6M from Andreessen Horowitz
Mem emerges from stealth. The pitch: "lightweight organization" - information summonable everywhere.
Series A: $23.5M led by the OpenAI Startup Fund
One of the fund's first investments. Post-money valuation: $110M. Total raised crosses $29M.
Mem 2.0 enters Alpha
A complete rebuild. Faster, more reliable, voice-first. The Year in Review captures the rollout.
Mem becomes an "AI Thought Partner"
The notes-app framing widens. A new pricing structure lands; Mem Pro arrives at $12 / month.
Voice, email, web, meetings - one inbox, no filing
Mem ships its full multi-modal capture set. The product is something your eight-folder-deep system never was: trusted.
By the dollar and the headcount
The financial story is small in venture terms and unusually clean. Two rounds. Two believers - one of which is now, plausibly, the most consequential AI company in the world. A team that has stayed deliberately compact while shipping a rewrite that would have broken larger companies.
Funding history, by round
Telepathy, eventually. Triage, immediately.
Ask Moody and Xu what they're actually building and the word that surfaces is "telepathy" - the idea that thoughts you've already had should be available to you, without searching, in the moment they're relevant. It is the kind of phrase that earns eye-rolls from skeptics and small, knowing nods from anyone who has ever stood in a meeting unable to recall their own opinion from two weeks ago.
Until the contact-lens version of the product ships - and the founders have, more than once, hinted at one - Mem's mission lives in the smaller, more boring victories. The forwarded email that becomes a tagged note without ceremony. The morning walk that turns into a usable meeting prep document. The chat with a customer that you didn't have to write down because Mem was listening, politely, and then filed itself.
The shape of the post-folder office
There is a version of the next decade in which everyone reading this still maintains a tangled filing system on their laptop. There is another version, faster arriving than most realize, in which capturing a thought becomes an act so light that you stop thinking about it - and the act of retrieving one becomes something close to instinctive. Mem is betting on the second version, and so far the bet has been more correct than it was reasonable to expect when the company started.
It is now that same Tuesday in 2026. Our product manager is back at her desk. She is preparing for a meeting with a customer she last spoke with in November. She doesn't open a folder. She doesn't search "John re: pricing." She walks toward the conference room and her notes - the recap from that prior call, the follow-up she meant to send, the philosophical aside about almonds (which, it turns out, the customer also mentioned) - are already waiting. She doesn't think about the system that put them there. Which is exactly the system she has, finally, learned to trust.