Hmmmm this sounds very familiar…

Google’s version if ICM?

The **Open Knowledge Format (OKF)** is an open, vendor-neutral specification designed to standardize how context, metadata, and curated knowledge are stored and shared between humans and AI agents.

Introduced by the Google Cloud Data Cloud team in June 2026, OKF formalizes the emerging **"LLM-wiki" pattern** (popularized by AI researcher Andrej Karpathy) into a portable, interoperable format. Instead of forcing AI agents to repeatedly search massive, disconnected text files or query proprietary vendor catalogs, OKF provides a shared standard for a living knowledge base.

### Core Technical Structure

The technical shape of an OKF "bundle" is intentionally minimal and lightweight. If you have ever used tools like Obsidian, Notion, or static site generators like Hugo, the structure will look highly familiar:

* **File System as Identity:** An OKF bundle is simply a directory tree of standard Markdown (.md) files. Each file represents a single unit of knowledge—called a **Concept**—which can be a physical asset (like a database table or API endpoint) or an abstract idea (like a business metric or an incident playbook).

* **YAML Frontmatter:** Every Markdown file begins with a snippet of YAML configuration metadata.

* **Minimal Schema Constraints:** To maintain ultimate flexibility, the specification enforces only **one** required field in the frontmatter: type. Other highly recommended (but optional) fields include title, description, resource (a unique URI pointing to the underlying asset), tags, and timestamp (ISO 8601).

* **Graph via Hyperlinks:** Concepts are linked together using standard Markdown links ([Link Text](path/to/file.md)). This effectively turns a flat folder structure into a rich graph of untyped, directed relationships (e.g., *parent/child*, *depends-on*, *joins-with*). Broken links are explicitly permitted to represent knowledge gaps that haven't been written yet.

### The 3 Core Design Principles

OKF was designed with strict guardrails to prevent it from turning into another bloated corporate data platform:

1. **Minimally Opinionated:** The format defines the *interoperability surface*, not the content model. Aside from requiring a type field, it does not mandate specific schema registries or structural sections.

2. **Producer/Consumer Independence:** It cleanly separates who writes the knowledge from who reads it. A bundle can be entirely hand-authored by a software engineer, generated automatically by an LLM parsing database schemas, or synthesized by an automated pipeline—and any AI agent or visualizer can read it seamlessly without translation.

3. **Format, Not Platform:** OKF is completely open-source and file-system-based. It requires no proprietary cloud architecture, database, model SDK, or runtime to execute. It can be version-controlled in a Git repository, compressed into a tarball, or hosted anywhere.

### Reference Implementations

To jumpstart adoption, the initial release includes open reference tools on GitHub:

* **The Enrichment Agent:** An agent that crawls a BigQuery dataset, automatically maps out OKF concept files for tables and views, and leverages an LLM to enrich those files with authoritative schema descriptions, citations, and potential data join paths.

* **The Static HTML Visualizer:** A single, self-contained HTML file with no backend dependencies. You drop an OKF bundle into it, and it renders a fully interactive graphical map of your organization's data context locally in the browser.

### Why It Matters

As enterprise AI shifts from simple retrieval-augmented generation (RAG) toward autonomous, agentic workflows, the primary bottleneck has shifted from raw model capability to **fragmented context**. OKF aims to serve as a universal *lingua franca* for systemic context, allowing developers to manage an AI's operational memory and corporate guardrails exactly like code.

1 comment