BundleDex / Guides / State of OKF 2026

State of OKF 2026

The definitive ecosystem report on the Open Knowledge Format — data-driven analysis of 829 bundles, 7068 stars, and the fastest-growing knowledge format for AI agents.

Updated 2026-10-04 ~15 min read

1. Executive Summary

The Open Knowledge Format (OKF), launched by Google Cloud Platform in June 2026, has seen explosive growth in its first weeks. BundleDex, the definitive OKF bundle directory, now indexes 829 bundles from 645 authors — accumulating 7068 total GitHub stars across 1122 unique tags. This report analyzes the ecosystem's composition, growth trajectory, and key trends.

Key Findings: The OKF ecosystem has grown from ~250 to 829 bundles in under a month. The top bundle (iwe) alone accounts for 25% of all stars. 312 bundles (38%) are OKF-conformant. The ecosystem added 4428 stars in 91 days.

2. Ecosystem Overview

829
Total Bundles
7,068
Total GitHub Stars
645
Unique Authors
1122
Unique Tags
1
Categories
312
OKF-Conformant

The OKF ecosystem is diverse, spanning AI agent skills, documentation tooling, domain-specific knowledge bases, and reference datasets. The format's simplicity — plain markdown files with YAML frontmatter — has lowered the barrier to entry, enabling contributions from individual developers, open-source projects, and organizations alike.

The average bundle has approximately 9 stars, though the distribution is heavily skewed by a few top-performers. The median bundle is newer and smaller, reflecting rapid ecosystem expansion.

3. Category Breakdown

Bundles span 1 categories. Here is the distribution:

Uncategorized
829 (100%)

Example Datasets and Uncategorized make up the largest cohorts, reflecting early experimentation. Content & Publishing (50 bundles) shows strong use of OKF for documentation-as-code, while Tooling & SDKs (36 bundles) demonstrates growing infrastructure around the format.

4. Top 10 Most-Starred Bundles

The following bundles have the highest GitHub star counts in the OKF ecosystem:

#1
iwe by iwe-org 1745 ⭐

Markdown memory system for you and your AI agent

ai-agentsai-memoryclihelixknowledge-graph +15 more
#2
okf-agent-memory by okf-memory 747 ⭐

Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zero external databases or dependencies. Built in pure Go.

agent-memoryai-agentsbm25coding-agentsdeveloper-tools +13 more
#3
lineage-skill by JuneYaooo 449 ⭐

Distill videos, PDFs, transcripts, and notes into source-backed teacher Agent Skills.

course-distillationeducation-aiknowledge-distillationknowledge-distillation-taskknowledge-management +5 more
#4
terrain by sopaco 251 ⭐

AI-native engineering environment management that makes your codebase agent-ready.

deepseekdeepwikidshdsh-pluginknowledge-base +2 more
#5
remnic by joshuaswarren 212 ⭐

Open-source memory and context for user-aware agents: scoped memory, provenance, retrieval quality, correction, boundaries, evals, and MCP/HTTP access.

ai-agentai-memoryclaude-codeengramhermes-plugin +14 more
#6
echoes-vault-opencode by psinetron 157 ⭐

Persistent memory plugin for OpenCode. Obsidian-style knowledge base that survives across sessions

agentic-memoryai-agentdeveloper-toolsknowledge-basellm +5 more
#7
Groma.md by MrLesk 152 ⭐

Your architecture as OKF Markdown in Git, and one C4 map you can walk. Scanned from source, curated by you and your agents.

agentsarchitecturebacklogbacklog-mdc4 +11 more
#8
cole-medin-ai-coding by coleam00 134 ⭐

An Open Knowledge Format (OKF) bundle of Cole Medin's best AI-coding videos — mount it in any AI second brain and search it deeply. Transcript-verified.

#9
cole-medin-knowledge-base by coleam00 132 ⭐

An OKF (Open Knowledge Format) knowledge base + Karpathy-style LLM wiki synthesized from Cole Medin's entire long-form YouTube catalog. Drop it next to any project for agent-ready, cited reference. No database, no embeddings.

#10
echoes-vault-codex by psinetron 131 ⭐

Persistent memory plugin for Codex. Obsidian-style knowledge base that survives across sessions

agentic-memoryagentic-memory-architectureai-agentcodexdeveloper-tools +7 more

5. Fastest-Growing Bundles

These bundles gained the most stars between 2026-07-05 and 2026-10-04 (absolute growth):

#1
iwe +471 ⭐

1274 → 1745 stars (37% growth) — Markdown memory system for you and your AI agent

#2
lineage-skill +182 ⭐

267 → 449 stars (68% growth) — Distill videos, PDFs, transcripts, and notes into source-backed teacher Agent Skills.

#3

60 → 134 stars (123% growth) — An Open Knowledge Format (OKF) bundle of Cole Medin's best AI-coding videos — mount it in any AI second brain and search it deeply. Transcript-verified.

#4
openknowledge +55 ⭐

7 → 62 stars (786% growth) — CLI tool for managing Open Knowledge Format (OKF) bundles.

#5
kiso +36 ⭐

11 → 47 stars (327% growth) — Kiso is a publishing engine that turns Open Knowledge Format (OKF) bundles into static websites for humans and AI agents.

Lineage Skill leads absolute growth with +471 stars, driven by its adoption as a Claude Code skill template. Data Olympus showed the highest percentage growth at 1500%, reflecting the growing interest in OKF-curated knowledge datasets.

6. Growth Trend

The OKF ecosystem has grown rapidly since launch. BundleDex initially tracked ~250 bundles. Today, we index 829 bundles — a 232% increase in just weeks. The star count grew from 2640 to 7,068 in 91 days alone.

Key Growth Driver: The launch of Claude Code native OKF support, Cursor integration, and OpenCode adoption have created strong demand for OKF bundles. Every major AI coding tool now supports the format, making it the de facto standard for agent knowledge packaging.

We expect continued acceleration as more developers discover OKF, more tools add native support, and the BundleDex discovery ecosystem (MCP server, JSON API, topic pages) makes finding the right bundle frictionless.

7. OKF Conformance

Of the 829 bundles indexed, 312 (38%) pass formal OKF conformance validation. This means they follow the exact directory structure, frontmatter schema, and cross-linking conventions defined in the OKF specification.

Non-conformant bundles may use OKF-like structures (markdown files with YAML frontmatter) but haven't been validated against the spec. We expect conformance rates to rise as validation tools like OKF Conformance and OKF Go Parser become more widely used.

→ See our guide on validating OKF bundles

8. Outlook & What's Next

The OKF ecosystem is still in its infancy. With 829 bundles and growing, we anticipate several trends:

  • Enterprise adoption: As OKF matures, expect enterprises to adopt it for internal knowledge management, replacing proprietary wiki solutions.
  • Specialized SDKs: More language-specific SDKs for loading and querying OKF bundles at runtime.
  • Cross-agent portability: OKF bundles that work across Claude Code, Cursor, OpenCode, and custom agent frameworks interchangeably.
  • Automated validation in CI: GitHub Actions and GitLab CI pipelines that validate OKF structure on every pull request.
  • AI-generated bundles: Tools that automatically convert existing documentation, wikis, and codebases into OKF format.

9. Frequently Asked Questions

How is this data collected?

BundleDex scans GitHub, GitLab, and other public repositories daily, indexing any repository tagged with 'okf' or meeting OKF structural criteria. Stars and metadata are refreshed on each scan.

What qualifies as an OKF bundle?

A repository that contains an index.md file with YAML frontmatter including okf_version, a concepts directory, and follows the OKF directory structure. Some bundles are indexed before full conformance is verified.

How often is this report updated?

This report is refreshed whenever the BundleDex index is regenerated. Check the trending page for real-time ecosystem metrics.

Explore the OKF Ecosystem

Browse all 829 bundles on BundleDex, or submit your own OKF bundle.

Browse All Bundles → Submit a Bundle →