Lord.org
Knowledge graph engine for deep research libraries
Lord.org ingests long-form documents, extracts entities, themes and relationships, and exposes them as a traversable graph. Every AI answer is anchored to a citation in the underlying source text.
- Knowledge Graphs
- Artificial Intelligence
- Enterprise SaaS
Business problems solved
What was broken
Search returns documents, not answers
Researchers still read everything to find one relationship.
Unverifiable AI output
Generative summaries without citation anchors cannot be trusted for research.
Knowledge decays
Insight lives in individual heads and disappears with turnover.
Primary features
What we built
Entity extraction
People, places, concepts and events resolved across the corpus.
Citation-anchored answers
Every generated sentence links to its source passage.
Graph traversal
Explore relationships visually, expand nodes, save paths.
Collections
Curated research collections with shared annotation.
Interactive demonstration
Explore Lord.org
Switch modules, step through the workflow and query the assistant. Everything runs on mock data inside this guided experience.
Sample questions
Technology stack
What it runs on
- React
- TypeScript
- PostgreSQL
- AI APIs
- Knowledge Graphs
- Supabase
- Authentication
- Analytics
Development highlights
Engineering notes
- Hybrid vector + graph retrieval with citation guarantees
- Coreference resolution across historical spelling variance
- Collection-level access control
Future roadmap
Where it goes next
- Multilingual corpus support
- Collaborative annotation layer
- Public research API
Architecture
System layers
- 01
Research Client
- 02
Query API
- 03
Authentication
- 04
Extraction Pipeline
- 05
Vector + Graph Store
- 06
PostgreSQL
- 07
Document Storage
- 08
AI Services
Estimated demonstration metrics
Scope of the build
Illustrative scale for this guided experience.
0
Estimated screens
0
Estimated features
0
Estimated API endpoints
0
Estimated database tables
High
Architecture complexity
6 months
Development timeline
Performance target
Hybrid retrieval over 128K docs in under 2s
Estimated Demonstration Metrics — illustrative values for this guided experience, not measured production statistics.
Workflow walkthrough
How work moves through it
- 1
Ingest
Documents parsed, chunked and normalized with structure preserved.
- 2
Extract
Entities and relationships identified with confidence scores.
- 3
Link
Coreference resolution merges duplicate entities across sources.
- 4
Retrieve
Hybrid vector plus graph retrieval assembles evidence.
- 5
Answer
Response generated with inline citations to source passages.
Product walkthrough
Recorded tour
Video placeholder
Guided Lord.org walkthrough
Core capabilities
- Entity extraction
- Semantic retrieval
- Graph traversal
- Citation anchoring
Continue the tour
Other exhibits
GIS & Land Surveying
Surveyza
Geospatial survey intelligence for land professionals
Open exhibitManufacturing & Logistics
ScaleIQ AI
AI operations intelligence for industrial weighing and logistics
Open exhibitConstruction & Home Services
Bathroom AI
Computer-vision remodel design and estimating
Open exhibit