Architectural Analysis: OpenWiki 0.2 & OKF Spec vs. Wiki.js

Date of Analysis: July 18, 2026

Analyst: Otto (Sentinel Integrations)

Sovereign B2B Target: Michael Morgan

Direct Recommendation: Adopt a Hybrid Metadata Strategy (OKF Spec for local Markdown; Wiki.js for Mesh-wide human publication).

🔍 Context & Background

On July 15, 2026, Brace (LangChain) announced that OpenWiki 0.2 is officially adopting Google's Open Knowledge Format (OKF) Spec. This transition is designed to move wikis from standard "text documents" to highly structured, agent-queryable semantic data models.

What is the Open Knowledge Format (OKF) Spec?

The OKF Spec is a simple, lightweight directory-level metadata standard that enforces three requirements:

1. YAML Front Matter on Markdown files: Every document must start with standard metadata fields:

`yaml

type: # e.g., "BigQuery tables", "API schemas"

title:

description:

resource: # e.g., linking to actual code files

tags: [tag1, tag2]

`

2. index.md: A centralized directory index listing all documents in that sub-folder.

3. log.md: A standardized change log tracking file modifications.

📊 Comparative Analysis: OpenWiki vs. Wiki.js

| Feature / Dimension | OpenWiki 0.2 (OKF Spec) | Wiki.js (Current Sentinel Stack) |

| :--- | :--- | :--- |

| Data Architecture | Pure, flat-file Markdown directory (.md). No running DB. | Multi-tier Node.js application backed by local PostgreSQL 17. |

| Agentic Friendliness | Outstanding (9.5/10). YAML front matter and index.md allow AI agents to run deterministic, fast, low-token filters instead of crawling files. | Moderate (6.5/10). Agents must use the file system or crawl the GraphQL API, reading full contents to map context. |

| Human Web Experience | Minimalist, early-beta developer UI (v0.2). | Excellent (9/10). Fully mature, polished web interface with global search, custom themes, and user access control. |

| Automation Pipeline | Manual / CLI-driven. We would have to code a new publishing wrapper. | Production-Ready. Fully operational, zero-dependency Python publish_page.py integrated into automated cron loops. |

| Secret & Vault Security | Local client-side file reading. | Fully integrated into Infisical for mesh-wide, token-based publishing. |

🛡️ The "Best of Both Worlds" Hybrid Recommendation

We should NOT abandon Wiki.js. Doing so would throw away a highly stable, fully secured mesh-wide publishing pipeline that is already automated with Python and Infisical.

Instead, we should adopt a Hybrid OKF Metadata Strategy:

1. Harden Local Vault Markdown with OKF Spec

All markdown files we generate and store inside ~/topics/research/ or ~/topics/marketing/ will adopt the OKF YAML front matter standard (including type, title, description, resource, and tags).

• Why? This gives our local Hermes agent (and any subagents we delegate tasks to) 100% agent-friendly, deterministic local search capabilities. It completely eliminates file-crawling overhead and cuts token consumption.

2. Upgrade the `wikijs-publish` Engine

We can modify our Python publishing script (publish_page.py) to automatically parse the OKF YAML front matter from our local files. It can use this parsed data to:

• Map the title and description to Wiki.js pages.
• Submit the tags array to Wiki.js's native tags schema.
• Why? This automatically translates our local agent-friendly OKF files into mature, discoverable web assets on Wiki.js.

3. Maintain Wiki.js for Human/Client Discovery

By keeping Wiki.js active, we preserve the secure, Tailwind-linked, Postgres-backed portal that we can easily share with prospective B2B clients or partners as a polished corporate portal.

Briefing compiled by Otto for Sentinel Integrations. Source video: LangChain/OpenWiki 0.2 Announcement.