#Google's New Open Knowledge Format Is Basically Basic Memory's Thesis

Google just formalized an idea Basic Memory has been built around from the beginning: useful AI context should live in plain files that people and agents can both read.

On June 12, Google Cloud announced the Open Knowledge Format, or OKF. It is an open specification for representing organizational knowledge as a directory of Markdown files with YAML frontmatter and links. What’s nice about it is that the goal is not some proprietary catalog or another vendor-owned memory silo. It is a portable shape for context: something humans can edit, agents can consume, and tools can move between.

It would be flattering to imagine Google arrived here because they noticed that Basic Memory has been doing the same thing since the introduction of MCPs. But the more interesting version is probably that the shape of AI memory is simply becoming obvious to everyone in the game.

AI agents need context, and context has to live somewhere. If it only lives inside one product, one vendor database, or one API, it is fragile. If it lives in ordinary files, people can read it, edit it, sync it, version it, move it, and keep it.

Google puts the point plainly:

“The answer to this problem isn’t another knowledge service. You need a format.”

Exactly.

That is what makes OKF noteworthy. It gives organizational knowledge a simple, portable form: Markdown, frontmatter, links, folders. An OKF bundle is a directory of Markdown files. Each file represents a concept: a dataset, a metric, an API, a decision, a process, or whatever else an agent might need to understand. Files can link to other files. index.md can help agents navigate. log.md can record history.

That is basically it.

And that is why it is interesting.

The old way of thinking about AI knowledge was: put everything into a system, then ask the system to be smart about it. OKF points toward a better pattern: keep the knowledge in a form humans and tools both understand, then let many systems use it.

Markdown. Frontmatter. Links. Files.

It is boring in the best possible sense. Like Git is boring. Like SQLite is boring. Like Markdown is boring. The kind of boring that outlives flashier approaches.

It also lines up closely with Basic Memory.

Same premise: memory shouldn’t disappear into someone else’s black box. It should live in ordinary Markdown files you control, readable by humans, writable by agents, synced when you want it, yours either way.

OKF gives that premise a name. But a frame isn’t a memory system.

A folder of Markdown files is a good start. Portable. Inspectable. Git-friendly. Already better than most AI memory products.

But once an agent actually has to use that folder, the questions start fast. How does it search across files? Follow the graph? Know what changed? Let a human and an agent edit the same knowledge without losing the structure? Show up in Claude, ChatGPT, Gemini, Cursor, Obsidian, everywhere you actually work?

That’s the part Basic Memory handles.

OKF is everyone agreeing to write knowledge on index cards instead of locking it in a filing cabinet. Basic Memory is the card catalog, the cross-references, the librarian, the circulation desk, everything that makes the cards useful every day.

So: did Basic Memory just become philosophically aligned with Google? Kind of. Mostly, Google just put a name on a pattern we’ve been building toward all along.

Either way — the future of AI memory isn’t a black box. It’s files: yours, readable, and useful everywhere you go.