vague v0.2.0-beta

BETA RELEASE v0.2.0 - OCR FEATURE

Local multimodal semantic search CLI. Search for images and text across your files with vague queries, OCR included for images.

Improvements from old version v0.1.1-beta:

• NEW FEATURE: OCR
  OCR allows you to extract text from images and save it inside the index, making the search for images full of text way easier. Can be used to find screenshots or graphs.
• Fixed bugs and stability issues.

Features included:

• Text and image search with CLIP embeddings
• PDF text extraction and indexing
• OCR scanning for images
• Hybrid scoring (text + image results ranked together)

What's in the Archive:

• vague.exe - The core CLI executable.
• models/ - The OCR model files bundled directly. While other models are automatically downloaded by vague setup, these specific OCR weights are included here to provide a seamless experience without requiring manual downloads due to limitations.

Notes & Limitations:

• The interface may change
• Only tested on Windows/Linux

HOW TO INSTALL & USE:
See README for installation and usage.

Feedback wanted:

• How well does it run?
• Feature suggestions
• Bug reports

Please open issues to report those! PRs from devs are welcome.
Fully written in Rust, 2nd minor version.