Vectify AI
To build reasoning-based document intelligence engines that achieve human-level accuracy on complex documents without relying on traditional vector databases.
At a Glance
- Enterprise Software
- FinTech
- LegalTech
- AI Infrastructure
AI Tools by Vectify AI
(1)
PageIndex
Vectorless RAG Framework for Devs
Discussions
No discussions yet
Be the first to start a discussion about Vectify AI
Latest News
Vectify AI Named Rising Star in Deep Tech by AlbionVC at SXSW London
PageIndex Adds Support for Anthropic and Ollama Models
Vectify AI's PageIndex Selected for GitHub Secure Open Source Fund
Perplexity Integrates PageIndex Concepts for Financial Vertical Analysis
Products & Services
An open-source, vectorless, reasoning-based RAG framework that uses document trees instead of embeddings.
Hosted version of PageIndex offering managed infrastructure, OCR capabilities, and agentic retrieval APIs.
A specialized reasoning-based model designed for high-accuracy financial question answering, achieving 98.7% on FinanceBench.
Market Position
PageIndex positions itself as a high-accuracy alternative to traditional vector-based RAG (e.g., LlamaIndex, LangChain) for complex, structured documents where precision is critical.
Leadership
Founders
Mingtian Zhang
CEO and Co-founder. PhD in Machine Learning from University College London (UCL) and Research Fellow at UCL's Centre for Artificial Intelligence. Expert in probabilistic modeling and RAG.
Yu Tang
Co-founder and Lead Developer. Researcher at UCL and the primary architect of the PageIndex framework. Specialized in agentic retrieval and document intelligence.
Executive Team
Mingtian Zhang
Chief Executive Officer
PhD in ML from UCL, Research Fellow at UCL Centre for AI.
Yu Tang
Co-founder & Lead Architect
Lead contributor to the PageIndex open-source project and document intelligence researcher.
Board of Directors
Founding Story
Vectify AI was founded by UCL researchers Mingtian Zhang and Yu Tang to solve the 'black box' and accuracy issues of vector-based RAG. They developed PageIndex to mimic how human experts navigate document structures.
Business Model
Revenue Model
Open-core model. Free and self-hostable open-source framework, with a hosted Cloud API and Chat API for enterprise/high-volume use.
Pricing Tiers
1,000 pages free on signup, access to basic API features.
Tiered pricing for high-volume document indexing and OCR-heavy workloads.
Managed deployments, custom integrations, and dedicated support.
Target Markets
- Enterprise Software
- FinTech
- LegalTech
- AI Infrastructure
- Financial analysis (10Q, 10K, annual reports)
- Legal document discovery and intelligence
- Medical and technical documentation QA
- High-accuracy enterprise search
- Perplexity
- UCL Centre for AI