Claude API
AI reasoning layer for document understanding, summarization, structured extraction, report generation, and workflow intelligence.
Infrastructure
Kaymai is being built with Claude as the intelligence layer and Cloudflare for application infrastructure — static delivery, edge execution, object storage, structured data, and AI observability.
The decided architecture behind Kaymai's AI workflows. Services that are still under evaluation are marked as such.
AI reasoning layer for document understanding, summarization, structured extraction, report generation, and workflow intelligence.
Backend APIs and workflow execution at the edge.
Object storage for source files and generated workflow outputs.
Structured metadata and application data for workflows.
Observability and management for AI requests.
Security and edge protection for Kaymai services.
Kaymai participates in the Cloudflare for Startups program.
Architecture flow
Workflow inputs move from the frontend through edge APIs and Claude-powered processing into storage, metadata, and dashboard review.
Step 1
Frontend
Step 2
Worker API
Step 3
Claude AI processing
Step 4
R2 storage
Step 5
D1 metadata
Step 6
Dashboard output
The public website is static today and deployable to Cloudflare Pages or Vercel without a backend service.
Product APIs use edge runtime patterns where they make sense for workflow execution.
The platform prioritizes protected uploads, controlled access, and clear operational boundaries as product scope expands.