Connectors
Data ingestion connectors for the Engram Context Engineering Platform.
Document Ingestion
The primary connector for enterprise knowledge is the Document Ingestion Pipeline:
- 📄 Document Ingestion Strategy - Full documentation with tri-search architecture
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:
- Keyword Layer - Zep sessions for BM25 search
- Vector Layer - pgvector embeddings for semantic search
- Graph Layer - Zep facts for knowledge graph queries