Alpha Plugin
nexus-knowledge-graph
v0.1.0AlphaTransforms uploaded documents into a searchable knowledge graph with entity and relationship extraction.
Overview
Uploaded PDFs, Markdown files, and DOCX documents are analyzed by an extraction service (AWS Textract) to identify entities, relationships, and structural elements. Results are stored in a PostgreSQL-native graph schema as part of the Nexus database, making them queryable by agents via dedicated MCP tools. Enables cross-document relationship discovery and semantic navigation.
Key Features
- PDF / Markdown / DOCX extraction pipeline
- Entity + relationship extraction (AWS Textract)
- PostgreSQL-native graph schema (part of Nexus DB)
- MCP tools for agent graph traversal
Registered Hooks
Events this plugin listens to in the OpenCode runtime.
document.ingestedTriggers extraction pipeline when a new document is uploaded or ingested into a project.
Custom Tools
MCP tools this plugin exposes to agents.
nexus_graph_queryTraverses the knowledge graph to find entities, relationships, or paths between concepts.
nexus_graph_statusReports extraction status and graph statistics for a project (node count, edge count, pending documents).