Skip to content

title:"Arkon: Enterprise AI Knowledge Hub & MCP Server"
tags:
- mcp-server
- ai-knowledge-base
- enterprise-ai
- rag-contexts
- departmental-workspaces

Arkon: Enterprise AI Knowledge Hub & MCP Server

Introduction

Arkon is an open-source, enterprise-grade platform designed to unify organizational knowledge with AI systems via the Model Context Protocol (MCP). It serves as a centralized MCP server, enabling teams to securely manage RAG (Retrieval-Augmented Generation) contexts, access policies, and AI skills. By bridging internal documentation, SOPs, and AI clients like Claude, Arkon eliminates fragmented AI adoption while ensuring traceable, role-scoped knowledge access.

Key Features

Intelligent Knowledge Wiki (MRP Pipeline)

Arkon’s MRP pipeline (Map → Reduce → Plan-review → Refine → Verify → Commit) compiles organizational documents into an interlinked wiki. This process ensures:
- Human review at each stage, with planners outlining page changes before execution.
- Content merging when sources overlap, preserving prior knowledge.
- Traceable claims, linking wiki pages to source documents.
- Image integration, embedding captions for visual context.
- Resumability, recovering from crashes without re-doing LLM processing.

Workspaces (Department & Project Scopes)

  • Departmental isolation: Separate knowledge wikis for HR, Legal, Engineering, etc.
  • Project-level scopes: Cross-functional teams or client projects with dedicated workspaces.
  • Strict scope enforcement: Access restricted at API, MCP, and search levels to prevent cross-context leaks.

Fine-Grained Role-Based Access Control (RBAC)

  • Built-in roles: Viewer, Contributor, Editor, Admin, and custom roles.
  • Granular permissions: Granular controls like wiki:edit:all or doc:read:own_dept.
  • Audit logs: Tracks privileged actions (e.g., role changes, plan approvals).

MCP Server for AI Clients

Arkon acts as an MCP server, allowing integration with Claude, Google Gemini, or OpenAI via OAuth 2.1 + PKCE. Exposed tools include:
- Wiki search and navigation (search_wiki, read_wiki_page).
- Source material tools (get_source, list_sources).
- Draft workflow (propose_wiki_edit, approve_draft).

Tech Stack

  • Backend: FastAPI, PostgreSQL + pgvector, Redis, MinIO.
  • Frontend: Next.js, Tailwind CSS.
  • AI Integration: Model Context Protocol (MCP), embeddings from Google or OpenAI.
  • Requirements: Docker, PostgreSQL, 4–16 GB RAM (scaling with team size).

Roadmap

  • Enhance MRP pipeline with deterministic compilation and crash recovery.
  • Expand workspaces with nested scopes and collaboration tools.
  • Introduce an Arkon CLI for quick deployment.
  • Add media ingestion support (videos, Excel files).
  • Build a usage analytics dashboard.

Conclusion

Arkon addresses the challenges of fragmented AI adoption by providing a secure, scalable knowledge management system. Its MCP server integration ensures AI tools like Claude leverage organizational data consistently and safely. With robust RBAC, modular architecture, and a focus on enterprise needs, Arkon is a compelling solution for teams seeking to institutionalize AI-driven workflows.

Original URL: https://github.com/nduckmink/arkon