A local-first, ownership-first knowledge repository designed for epistemic durability.
"Treat knowledge as durable infrastructure, not disposable notes."
The Knowledge system is a WordPress plugin built for individuals and organizations who need to retain, analyze, and reason over information long-term. It solves the problem of fragile knowledge—link rot, silent content changes, and opaque AI answers—by enforcing strict ownership and provenance.
Unlike standard note-taking apps or "Second Brain" tools, this system prioritizes trust and correctness over speed or convenience.
- Immutable Truth: When you ingest a source (URL, PDF, Text), it creates a Version that is never overwritten.
- Controlled Divergence: If you want to edit or reinterpret content, you create a Fork. You never silently rewrite history.
- Local-First: Content is stored in a deterministic filesystem layout (
/kb-data/), not just the database. It works offline. - AI as Servant: AI (LLMs, RAG) is used to summarize and reason, but it is never authoritative. It reads from your knowledge base but cannot write to it without your explicit approval.
- Robust Ingestion: Capture full HTML/Text from URLs and Documents.
- Deep Versioning: Keep the original source forever. Compare versions over time.
- Active Reading & Annotation: Highlight text and add notes in the sidebar. Notes are anchored to immutable text versions.
- Projects & Scoped Context: Organize knowledge into Projects. Run Search and RAG queries scoped only to a specific project's context.
- Local AI / RAG: Run RAG (Retrieval Augmented Generation) queries against your data using local models (Ollama) or cloud providers (OpenAI).
- Multi-Provider AI: Configure failover chains for AI providers (e.g., try Local Ollama first, failover to OpenAI).
- Elementor Integration: Dedicated "Knowledge Archive" and "Knowledge Search" widgets with advanced controls (Infinite Scroll, Note Indicators, Results Dividers).
- Mobile Optimized: Fully responsive Admin UI and Frontend for touch devices.
- Filesystem Backed: All content and media are stored as plain files. The database is just an index.
- Platform: WordPress (PHP)
- Storage:
- Metadata: WordPress Database (Custom Post Types:
kb_article,kb_version,kb_fork,kb_project,kb_note) - Content: Local Filesystem (
/wp-content/kb-data/) - Search: Custom Shadow Index (
wp_kb_search_index) + Local Vector Store
- Metadata: WordPress Database (Custom Post Types:
- AI Integration: HTTP API to local inference server (e.g., Ollama) or OpenAI API.
- WordPress 6.0+
- PHP 8.0+
- Write access to
/wp-content/(for creating thekb-datadirectory) - (Optional) Ollama running locally for local AI features.
- (Optional) OpenAI API Key for cloud-based AI features.
- Clone this repository into your
wp-content/plugins/directory.cd wp-content/plugins git clone https://github.com/stevecopeza/knowledge.git - Activate the plugin in WordPress Admin.
- The system will automatically create the
/wp-content/kb-data/directory structure.
Detailed architectural specifications can be found in the docs/ directory:
- Executive Summary
- Domain Model
- Storage Spec
- Annotations & Notes
- Projects & Scope
- Search & RAG
- AI Integration
- Elementor Integration
This project is currently in Beta / MVP Phase. We prioritize correctness over features.
Proprietary / Internal Use (TBD)