Locus v1.0.0

Locus v1.0.0

This is the first stable release after pre‑release builds.

Compared to the pre-release, this version adds:

• Chunked indexing for more precise results
• Optional OCR for scanned PDFs (RapidOCR/ONNX) with caching
• RRF rank fusion (default) for more stable relevance
• Rebuild Index detection when PDFs or models change
• Cancel indexing and clear cache actions in Manage Models
• Cleaner UI labels and an icon
• Many small bugs fixed

Thanks for trying Locus!

Locus v0.3.0 - chinese version

Support Chinese as the system and pdf language, while keeping all the previous features.
添加了中文，并优化了中文pdf的处理。所有原功能保持不变。

Locus v0.2.0: Faster & Lighter

This release transforms Locus into a lightweight, high-performance search engine by migrating from heavy PyTorch-based models to FastEmbed (ONNX). We have streamlined the core engine and GUI to ensure a smaller disk footprint and faster initialization.

Locus v0.1.1: Local Hybrid PDF Search Engine

📚 Locus – Offline PDF Search Tool

Locus is a lightweight, offline desktop tool designed to help students and researchers find the exact page that answers their questions within a library of PDFs. By combining BM25 keyword matching with semantic similarity, Locus understands both the exact words you type and the meaning behind your questions.

---

✨ Key Highlights

• Hybrid Search
  Toggle between Semantic mode (for conceptual understanding) and Literal mode (for exact term matching) using our custom UI slider.

• Offline Privacy
  Your documents never leave your machine—100% local processing with no cloud dependencies.

• Jump to Page
  Double-click any result to open your PDF directly at the relevant page (optimized for SumatraPDF).

• Hardware Optimized
  Choose from four quality profiles (Fast to Best) to match your PC's RAM and CPU capabilities.

---

🚀 How to Run

1. Download Locus-v0.1.1.zip attached to this release.
2. Extract the files from the zip.
3. Double-click Locus.exe to launch the application.

⚠️ Note: The first launch requires an internet connection to download required models (~80MB to ~6GB). After setup, Locus works completely offline.

---

🛡️ Security & Verification

Since this application is not digitally signed, Windows SmartScreen may display a warning. You can verify file integrity using the SHA-256 hash below:

SHA-256 Hash of Locus-v0.1.1.zip:
B0E8CD41B5397C9EEE0AC6B2DBCEBF24554439E0A0110143A11C611F4235A1D0

🛠️ System Requirements

• Disk Space: ~2.5 GB
• RAM: 4 GB minimum (8 GB+ recommended for "High Accuracy" mode)
• PDF Viewer: For the "Open at Page" feature to work, we highly recommend installing SumatraPDF

Locus v0.1.0-beta.1: Local Hybrid PDF Search Engine

📚 Locus – Offline PDF Search Tool

Locus is a lightweight, offline desktop tool designed to help students and researchers find the exact page that answers their questions within a library of PDFs. By combining BM25 keyword matching with semantic similarity, Locus understands both the exact words you type and the meaning behind your questions.

---

✨ Key Highlights

• Hybrid Search
  Toggle between Semantic mode (for conceptual understanding) and Literal mode (for exact term matching) using our custom UI slider.

• Offline Privacy
  Your documents never leave your machine—100% local processing with no cloud dependencies.

• Jump to Page
  Double-click any result to open your PDF directly at the relevant page (optimized for SumatraPDF).

• Hardware Optimized
  Choose from four quality profiles (Fast to Best) to match your PC's RAM and CPU capabilities.

---

🚀 How to Run

1. Download the Locus.exe file attached to this release.
2. Double-click Locus.exe to launch the application.

⚠️ Note: The first launch requires an internet connection to download required models (~20-800 MB). After setup, Locus works completely offline.

---

🛡️ Security & Verification

Since this application is not digitally signed, Windows SmartScreen may display a warning. You can verify file integrity using the SHA-256 hash below:

SHA-256 Hash:
3362DCA718C017732EA27C9D106C71242678881A1B47E3EAA625AB041FE83233

🛠️ System Requirements

• Disk Space: ~800 MB (includes pre-packaged transformer models for offline use)
• RAM: 4 GB minimum (8 GB+ recommended for "High Accuracy" mode)
• PDF Viewer: For the "Open at Page" feature to work, we highly recommend installing SumatraPDF