Show HN: Graphora – Turn text documents into graph databases
2 ezhil 4 6/10/2025, 11:13:33 AM
Hi HN,
Built this after manually creating graph databases took me a few months at my last company.
Graphora converts unstructured text (documents, emails, reports) into graph databases automatically:
- Upload text → LLM extracts entities & relationships
- Handles entity resolution (same person, different names)
- Interactive ontology designer
- Exports to Neo4j (for now)
Demo: https://demo.graphora.io
You'll need to provide:
- Gemini API key (your own)
- 2 Neo4j instances for staging & Prod. [You can sign up for Free Aura]
All credentials encrypted.
Looking for feedback from anyone who's wrestled with turning messy text into structured relationship data. Let us know!
Comments (4)
Oras · 19h ago
How does it work when uploading documents later? let's say I uploaded a batch of 50 documents, then a week later I uploaded another 20. How does it ensure the correlation and context?
ezhil · 19h ago
The extraction happens based on a schema of nodes and edges. So, let's say you have 2 docs that have related data on a company, then they both will be connected. We use Entity resolution to combine them.
Oras · 19h ago
So you need a pre-defined schema? and then when uploading, you assign which schema to use?
ezhil · 18h ago
Yes manually assigned for now. Future versions will have detection based on docs