Connectors

Data ingestion connectors for the Engram Context Engineering Platform.

Document Ingestion

The primary connector for enterprise knowledge is the Document Ingestion Pipeline:

Connector Types

Connector Status Description
File Upload ✅ Implemented PDF, DOCX, TXT, MD via /api/v1/etl/ingest
SharePoint 🔶 Planned Enterprise document libraries
Confluence 🔶 Planned Wiki pages and spaces
GitHub 🔶 Planned Source code and documentation
ServiceNow 🔶 Planned Tickets and incidents

Architecture

Connectors feed into the tri-indexing pipeline:

  1. Keyword Layer - Zep sessions for BM25 search
  2. Vector Layer - pgvector embeddings for semantic search
  3. Graph Layer - Zep facts for knowledge graph queries