<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/using-okf-to-provide-long-term-context-to-your-agent-9rqtpfok6" -->

---
title: Using OKF to provide long term context to your agent
description: An engineer describes adopting the Open Knowledge Format (OKF), an open spec from Google Cloud's Knowledge Catalog team, to give AI coding agents persistent,...
canonical: https://daily.dev/posts/using-okf-to-provide-long-term-context-to-your-agent-9rqtpfok6
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: Using OKF to provide long term context to your agent | daily.dev
og:description: An engineer describes adopting the Open Knowledge Format (OKF), an open spec from Google Cloud's Knowledge Catalog team, to give AI coding agents persistent,...
og:url: https://daily.dev/posts/using-okf-to-provide-long-term-context-to-your-agent-9rqtpfok6
og:image: https://api.daily.dev/og/posts/9RqtpfOK6.png
og:image:alt: Using OKF to provide long term context to your agent
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Using OKF to provide long term context to your agent

**[iO tech\_hub](https://daily.dev/sources/iotechhub)** · 17 min read · 10 upvotes · 0 comments

## Summary

An engineer describes adopting the Open Knowledge Format (OKF), an open spec from Google Cloud's Knowledge Catalog team, to give AI coding agents persistent, repo-versioned memory instead of relying on tool-specific memory files hidden under a user's home folder. OKF is a directory of markdown files with YAML front matter where one file equals one concept, paths are stable identifiers, and per-directory index.md and log.md files act as a catalogue and changelog. Version 0.2 adds provenance and trust fields (sources, generated, verified, status, stale_after) so a document's trustworthiness can be judged before reading it. The piece contrasts OKF with RAG/embeddings (which re-derive answers each time) and MCP (a transport protocol that can serve an OKF bundle), argues design decisions and subsystem pages are the highest-value content to keep, and explains how OKF complements Spec Driven Development by separating disposable workflow-shaped plans from durable state-shaped designs. Includes a minimal directory setup, an AGENTS.md snippet to wire agents to read and update the bundle, and a linked example repository (okf-drift).

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://techhub.iodigital.com/articles/using-okf-to-provide-long-term-context-to-your-agent/using-okf-to-provide-long-term-context-to-your-agent>

## Questions this post answers

### What is the Open Knowledge Format (OKF) and what fields does it require?

OKF is an open specification, published by the Google Cloud team behind their Knowledge Catalog, for storing knowledge as a directory of markdown files with YAML front matter. Only the type field is required and it is a free string. Each file represents one concept, and its path serves as that concept's identity, so moving a file changes its id.

_Teams standardizing agent memory formats can track spec updates like OKF through daily.dev._

### What new fields did OKF version 0.2 add compared to version 0.1?

Version 0.2 added provenance and trust fields on top of version 0.1's type, title, description, resource, and tags fields. The new fields are sources, generated, verified, status, and stale_after, which let a consumer judge whether a document is trustworthy and current before spending tokens reading it. Actors are written as producer/version for tools, human:id for people, or process:id for automation.

_Anyone deciding how to structure agent-readable docs can follow format changes like this via daily.dev._

### How does a markdown knowledge base like OKF differ from RAG or MCP for AI agent context?

Retrieval-augmented generation re-derives an answer from raw source chunks on every query and discards the synthesis afterward, so nothing accumulates over time. A markdown wiki like OKF is a compounding artifact where cross-references and prior contradictions stay resolved once written. MCP is a separate transport protocol for reaching tools or systems, and it can serve an OKF bundle rather than compete with it.

_Developers weighing RAG, MCP, and persistent docs for agent memory can follow the comparison on daily.dev._

## Similar posts on daily.dev

- [Knowledge as Code: The Memory File Just Got a Spec](https://daily.dev/posts/knowledge-as-code-the-memory-file-just-got-a-spec-rlawyv7rd) · Pulumi · 2 upvotes · 0 comments
- [How the Open Knowledge Format can improve data sharing](https://daily.dev/posts/how-the-open-knowledge-format-can-improve-data-sharing-qqswihmau) · Google Cloud · 15 upvotes · 0 comments
- [Architecture as Code: How Google’s OKF Makes Coding Agents Follow Your Architecture](https://daily.dev/posts/architecture-as-code-how-google-s-okf-makes-coding-agents-follow-your-architecture-cc2o3991u) · Medium · 2 upvotes · 0 comments
- [Google’s OKF: The Missing Layer Between RAG and AI Agents](https://daily.dev/posts/google-s-okf-the-missing-layer-between-rag-and-ai-agents-ifmrkg5ll) · Medium · 2 upvotes · 0 comments

---

Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#claude-code](https://daily.dev/tags/claude-code), [#context-engineering](https://daily.dev/tags/context-engineering)

[View this post on daily.dev](https://daily.dev/posts/using-okf-to-provide-long-term-context-to-your-agent-9rqtpfok6)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"Using OKF to provide long term context to your agent","url":"https://daily.dev/posts/using-okf-to-provide-long-term-context-to-your-agent-9rqtpfok6","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/using-okf-to-provide-long-term-context-to-your-agent-9rqtpfok6"},"datePublished":"2026-09-11T11:43:08.492Z","dateModified":"2026-10-15T02:22:20.686Z","description":"An engineer describes adopting the Open Knowledge Format (OKF), an open spec from Google Cloud's Knowledge Catalog team, to give AI coding agents persistent,...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/c3ffeaebb52da5c732100290198b991a?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/c3ffeaebb52da5c732100290198b991a?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"iO tech_hub","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"iO tech_hub","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/ac540f589d8843d99692bd6368ee7680","url":"https://daily.dev/sources/iotechhub"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/using-okf-to-provide-long-term-context-to-your-agent-9rqtpfok6","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":10},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,claude-code,context-engineering","timeRequired":"PT17M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"iO tech_hub","item":"https://daily.dev/sources/iotechhub"},{"@type":"ListItem","position":3,"name":"Using OKF to provide long term context to your agent"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/using-okf-to-provide-long-term-context-to-your-agent-9rqtpfok6#faq","mainEntity":[{"@type":"Question","name":"What is the Open Knowledge Format (OKF) and what fields does it require?","acceptedAnswer":{"@type":"Answer","text":"OKF is an open specification, published by the Google Cloud team behind their Knowledge Catalog, for storing knowledge as a directory of markdown files with YAML front matter. Only the type field is required and it is a free string. Each file represents one concept, and its path serves as that concept's identity, so moving a file changes its id. Teams standardizing agent memory formats can track spec updates like OKF through daily.dev."}},{"@type":"Question","name":"What new fields did OKF version 0.2 add compared to version 0.1?","acceptedAnswer":{"@type":"Answer","text":"Version 0.2 added provenance and trust fields on top of version 0.1's type, title, description, resource, and tags fields. The new fields are sources, generated, verified, status, and stale_after, which let a consumer judge whether a document is trustworthy and current before spending tokens reading it. Actors are written as producer/version for tools, human:id for people, or process:id for automation. Anyone deciding how to structure agent-readable docs can follow format changes like this via daily.dev."}},{"@type":"Question","name":"How does a markdown knowledge base like OKF differ from RAG or MCP for AI agent context?","acceptedAnswer":{"@type":"Answer","text":"Retrieval-augmented generation re-derives an answer from raw source chunks on every query and discards the synthesis afterward, so nothing accumulates over time. A markdown wiki like OKF is a compounding artifact where cross-references and prior contradictions stay resolved once written. MCP is a separate transport protocol for reaching tools or systems, and it can serve an OKF bundle rather than compete with it. Developers weighing RAG, MCP, and persistent docs for agent memory can follow the comparison on daily.dev."}}]}
```

