QuoteHub

Quotation document processing system. Upload PDF or XLSX quotations, extract structured data, and search across all processed documents.

Version: v0.065.0 — shown in the app header.

Features

• AI-powered extraction for scanned PDFs, text PDFs, and XLSX files
• Multi-page processing with streaming progress
• Full-text search with prefix matching
• Editable results with find-and-replace
• Encrypted backup/restore (AES-256-GCM .quodb packages)
• Supplier database — contacts, aliases, brands, scan, merge, purge
• File-at-rest encryption (AES-256-GCM, optional)
• Auto-backup with daily/weekly retention
• Three-role authentication (user / admin / master)
• CLI for pre-update backups and key management

Prerequisites

• Docker — Install Docker
• AI Server — any OpenAI-compatible VLM API (LM Studio, vLLM, etc.)

Quick Start

1. Clone and configure

git clone https://github.com/Electrasia/quotehub.git
cd quotehub
cp config.example.json config.json
# Edit config.json with your AI server details, or fill them in via Settings

2. Build and run

export FILE_ENCRYPTION_KEY="$(python3 -c 'import os;print(os.urandom(32).hex())')"  # optional
./deploy.sh

3. Open QuoteHub

Open http://localhost:8000 in your browser.

4. Connect AI Server

Go to Settings → enter your AI endpoint and model → Save Settings → Connect to AI Server in the header.

Updating

./deploy.sh

Auto-detects fresh install vs update. Creates a pre-update backup automatically when updating.

Authentication

Role	Search	Upload / Process / Edit	AI Settings	Export / Import	User Mgmt
user	✓	—	—	—	—
admin	✓	✓	view only	—	—
master	✓	✓	✓	✓	✓

On first startup, a master user is created with a random password printed to the logs. Change it immediately.

Configuration

config.json is mounted from the host and persists across rebuilds.

Setting	Description	Default
ai_endpoint	AI server URL	""
model	Model name	""
timeout	Request timeout (seconds)	90
max_retries	Max retry attempts	2
external_url	QuoteHub URL for image access	""
popup_duration	Success popup duration (seconds)	3
ocr_enabled	Enable OCR for scanned PDFs	true
extraction_enabled	Enable AI extraction	true
max_upload_size_mb	Max upload file size (1–20 MB)	5
trust_proxy_headers	Trust proxy headers (NPM)	false

Backup & Restore

All exports use AES-256-GCM encrypted .quodb packages. Export/import are master-only.

• Manual: Settings → Backup / Restore → Export / Import
• Automatic: Daily at 03:00, weekly promotion on Sundays, event-based before imports and updates

Project Structure

quotehub/
├── backend/              Python, FastAPI, SQLite
├── frontend/             Vanilla HTML/CSS/JS
├── data/                 Runtime data (Docker volume)
├── config.json           Personal settings (mounted)
├── deploy.sh             One-command deploy/update
├── VERSION               Current release version
├── Dockerfile            Docker image
└── docker-compose.yml    Docker Compose config

Tech Stack

• Backend: Python, FastAPI, SQLite (FTS5), httpx
• Frontend: Vanilla HTML/CSS/JavaScript (no frameworks)
• AI: Any OpenAI-compatible VLM API
• Container: Docker with Python 3.11-slim, non-root user
• Encryption: AES-256-GCM at rest + for backup packages

Troubleshooting

• Container not healthy: docker compose logs quodb --tail 50
• Reset and redeploy: docker compose down && ./deploy.sh
• Permission errors: ensure Docker has access to the data volume
• Health check: curl http://localhost:8000/health

License

MIT