Skip to content

📦 Montscan v1.3.0

Choose a tag to compare

View all tags
@sh4den sh4den released this 15 Jun 22:47
· 7 commits to main since this release
v1.3.0
af16055

What's New

Folder-Based Ingress 📂

Montscan can now watch a local directory for incoming PDFs instead of (or alongside) FTP and Samba. Drop a scanned document into a folder and Montscan will rename it with an AI-generated name — either in place or by moving it to a separate output folder.

Configurable Ollama GPU Acceleration ⚡

Ollama in Docker Compose can now run on CPU or NVIDIA GPU, switchable entirely from .env — no more editing docker-compose.yml by hand.

Changes

  • Folder Watcher Ingress: New FOLDER_ENABLED, FOLDER_INPUT_DIR, FOLDER_OUTPUT_DIR, and FOLDER_POLL_INTERVAL_SEC settings to monitor a local directory for new PDFs
  • Rename-in-Place or Move-to-Output: Omit FOLDER_OUTPUT_DIR to rename files where they sit, or set it to move renamed files to a separate output directory
  • Docker GPU Toggle: New COMPOSE_PROFILES and OLLAMA_GPU_COUNT variables select between CPU-only and GPU-accelerated Ollama containers
  • Documentation Updates: README now covers folder watcher setups and GPU configuration

Configuration

Enable the folder watcher:

FOLDER_ENABLED=true
FOLDER_INPUT_DIR=./input
FOLDER_OUTPUT_DIR=./output   # omit to rename in place
FOLDER_POLL_INTERVAL_SEC=5

Enable GPU acceleration for Ollama in Docker Compose:

COMPOSE_PROFILES=gpu
OLLAMA_GPU_COUNT=1

Leave COMPOSE_PROFILES=cpu (or unset) for CPU-only.

Technical Details

  • Added server/folder.go for the folder watcher loop and providers/local.go for rename/move handling
  • Updated config/config.go and main.go to wire up the new folder ingress
  • Updated agent/agent.go to support the local provider
  • Split the Ollama service in docker-compose.yml into ollama (CPU) and ollama-gpu (NVIDIA) variants behind Compose profiles, sharing config via a YAML anchor

Upgrade Notes

The folder watcher is disabled by default (FOLDER_ENABLED=false); existing FTP/Samba setups are unaffected. Ollama continues to run on CPU by default in Docker Compose unless COMPOSE_PROFILES=gpu is set.

Full Changelog: View commits since v1.2.0

Assets 2
Loading