Description
OpenCrab is a knowledge graph platform that turns documents, web research, and domain knowledge into reusable ontology packs and graph RAG workflows. It converts scattered knowledge into an executable structure, transforming data into ontology packs, packs into skills, and skills into repeatable, specialized agents. The core workflow moves from ingestion to graph RAG: users ingest documents, URLs, notes, and structured research; OpenCrab extracts entities, concepts, evidence, relationships, and claims; it stores them as an ontology graph organized into a multi-space meta-ontology; and users then query that graph from OpenCrab, MCP clients, and connected AI workspaces.
Because OpenCrab exposes an MCP endpoint, the same knowledge can be connected to external AI tools such as ChatGPT, Claude, Claude Code, Codex, Cursor, and Antigravity, giving agents a single grounded endpoint with source-verifiable, evidence-based answers. Users can build private, isolated graph RAG workspaces from their own material, and domain experts can package expertise into ontology packs to publish and sell through a marketplace after a quality review.
OpenCrab is aimed at researchers and analysts building personal evidence graphs, domain experts turning expertise into sellable packs, and AI builders grounding agents with connected knowledge. Use cases include marketing and content planning, brand and customer analysis, research synthesis, and building private knowledge-based graph RAG.
Pricing includes a free plan to explore public ontology packs and query shared graphs, a Pro plan for ingesting private documents into an isolated personal workspace, an Expert plan for creating, managing, and selling reviewed ontology packs with advanced reverse ingestion, and an Enterprise plan for teams with a dedicated ontology workspace and organization-level permissions.
OpenCrab's Core Features
Turn documents, URLs, and research into reusable ontology packs
Graph RAG workflow from ingestion to query
Entity, concept, evidence, and relationship extraction
Ontology graph organized into a multi-space meta-ontology
Single MCP endpoint for ChatGPT, Claude, Codex, Cursor, and more
Private, isolated graph RAG workspaces
Marketplace to publish and sell reviewed ontology packs
Source-verifiable, evidence-based answers
How to use OpenCrab?
Sign up: Create an account and open the dashboard with Marketplace, Packs, Ingest, and MCP menus.
Choose a plan: Pick Free to use public packs, or Pro and above to ingest your own material.
Ingest knowledge: Add documents, URLs, notes, and structured research to build an evidence graph.
Generate an MCP URL: Create an endpoint and register it in AI tools like ChatGPT, Claude, or Codex.
Query the graph: Ask questions and reason over your ontology graph from connected AI workspaces.
OpenCrab's Use Cases
- Research evidence graphs
- Sellable knowledge packs
- Grounding AI agents
- Content and analysis
