Cinnamon AI
Tokyo-based AI company offering Flax Scanner AI-OCR platform and specialized business AI agents for financial appraisal and knowledge management.
Overview
Founded in 2012 in Tokyo, Cinnamon AI has evolved from document processing to specialized business AI agents. The company serves over 50 enterprise customers including Toyota, Sumitomo Pharma, Fujitsu, and Daikin across four continents with approximately 119 employees. After raising $39M in Series C funding in April 2020, the company has expanded beyond its core Flax Scanner OCR platform.
In 2026, CEO Miki Hirano outlined the company's strategic shift toward "process redesign" rather than general AI distribution, emphasizing specialized business agents with limited objectives. Hirano's appointments to Japan's Growth Strategy Council and AI and Semiconductor Working Group position Cinnamon AI within national AI policy development.
The company now deploys Financial Appraisal AI Agents for automated first and second appraisal processes in financial services, and self-improving RAG systems for technical documentation research, signaling expansion beyond document extraction into knowledge management.
Key Features
- Flax Scanner: Coordinate-free AI-OCR using feature learning technology requiring no template creation
- Flax Scanner HUB: Generative AI-powered OCR with context analysis and no-code learning
- Financial Appraisal AI Agents: Automated appraisal processes with human oversight for later-stage work
- Super RAG: Self-improving retrieval systems integrated with AI agents for technical documentation
- Multi-Format Processing: Handles industry documents, technical drawings, and photographic images
- 3-Second Processing: Document processing speed of approximately 3 seconds per sheet
Use Cases
Financial Services Automation
Financial institutions deploy Cinnamon AI's appraisal agents for automated first and second appraisal processes, with human oversight reserved for complex later-stage evaluations.
Invoice Processing
Finance teams use Flax Scanner achieving 92.99% character accuracy and 86.17% item-specific accuracy across 27 invoice items without template configuration.
Technical Documentation Research
Engineering teams implement self-improving RAG systems that learn from technical documentation queries, enhancing knowledge management workflows.
Trade Document Processing
Logistics departments process shipping documents with 91% accuracy across 54 Commercial Invoice items and automatic extraction of 50 Bill of Lading items.
Technical Specifications
| Feature | Specification |
|---|---|
| Core Products | Flax Scanner, Flax Scanner HUB, Financial Appraisal AI Agents, Super RAG |
| Technology | Feature learning AI-OCR, generative AI, self-improving RAG |
| Template Requirements | None (coordinate-free extraction) |
| Processing Speed | ~3 seconds per document |
| Invoice Accuracy | 92.99% character, 86.17% item-specific |
| Trade Document Accuracy | 91% (Commercial Invoice, 54 items) |
| Deployment Options | Multi-tenant SaaS, single tenant, on-premise, private cloud |
| Integration | API support |
| Learning Function | No-code semi-automatic learning |
Resources
- Website
- Flax Scanner Product Page
- Flax Scanner HUB Platform
- AI Trends 2026 Analysis
- Pitch Video 1
- Pitch Video 2
Company Information
Headquarters: Tokyo, Japan (offices in US, Vietnam, Taiwan)
Founded: 2012
Founders: Miku Hirano (CEO), Hajime Hotta, Hiroaki Kitano, Mori Aki, Yoshiaki Ieda
Employees: ~119 (across 4 continents)
Funding: $39M Series C (April 2020), $52.9M total raised