SheetSense
Self-hosted Ollama-powered document text extraction with web UI
This solution provides a self-hosted OCR service using Ollama's open-source models, deployed on a Proxmox LXC container with GPU acceleration for fast processing. A scalable reverse proxy setup allows multiple clients to access the API-based OCR functionality securely, enabling instant results via HTTP requests. The system includes an intuitive OpenWebUI frontend, which can be hosted on your company website to let users interact with the OCR service directly. The architecture supports multi-user access and leverages Proxmox's flexibility for easy scaling.
Role
Server Admin, Scale Engineer
Frontend stack
OpenWebUI
Open source self-hosted UI
Data & APIs
Ollama API
Encoded image data
Deployment
LXC Containers
Bash scripts
01Drag-and-drop image upload with real-time preview and validation capabilities.
02Seamless API integration with self-hosted Ollama for intelligent text extraction.
03Extracted text display, editing, and export workflows with batch processing support.
04Light mode.
05Dark mode.