Proof of Usefulness Score: 85
---
DeepSearch is a promising cross-platform file indexing utility leveraging Rust for high-performance searches, particularly on SMB shares. While it demonstrates strong technical innovation and a clear problem-solution fit, the project is extremely early-stage and currently lacks verifiable user adoption, significant audience reach, or established market traction.
What is DeepSearch?
Scrolling and indexing software => extremely fast file searching by name
It was created to solve a problem: extremely slow file searching on shared SMB drives.
Who is DeepSearch for?
- Ordinary users who want to speed up file searches
- Security professionals who scan entire systems
- Hackers scanning files to find configuration files, password files, and user keys.
DeepSearch's Competitor: Voidtools' Everything
- Everything scans extremely fast due to indexing via loading MFT files (a type of log file system for the NTFS Windows file system).
- Everything is free; but it's closed source.
So what advantages does DeepSearch have over Everything?
1. DeepSearch runs on multiple platforms: Windows, Linux, MacOS... Because DeepSearch is written in Rust and its libraries, it doesn't depend on the operating system and leverages the cross-platform capabilities of the Rust ecosystem. Everything is only available on Windows.
2. DeepSearch is memory-safe: DeepSearch is written in Rust, so it achieves absolute memory safety. Any memory leak errors are detected and forced to be handled during compilation. It's not as potentially dangerous as Everything's C++.
3. Speed, speed, and speed: DeepSearch doesn't rely on MFTs; it's accelerated by a multi-tiered scanning strategy and Rayon's multi-threaded optimization engine. DeepSearch's scanning algorithm divides the scanning process into two stages:
Stage 1: The scan indexes all files and directories at the first level to provide "entry points" for Rayon threads in Stage 2.
Stage 2: Rayon threads are divided and enter the "entry points" obtained from Stage 1. They recursively scan the directories together.
The indexes are stored in the built-in "redb" database. Search results by file name are complete and returned almost immediately. When "rescanning," DeepSearch does not immediately overwrite the old index. Instead, it swaps the old version for the new version after the "rescan" process is complete. This ensures both safety and allows the search operation to continue while the "rescan" process is running in the background.
Therefore, DeepSearch can both speed up indexing on local drives and speed up scanning of network drives (SMB/Share), as long as the operating system can access them. This strategy is effective even in noisy environments like Wi-Fi. This is an area where Everything is still lacking. Everything is very fast on local drives. But on network drives, Everything is extremely slow due to sequential browsing. To be faster, Everything requires the server to deploy a "server" version of itself.
4. Instant search without indexing: Even without scanning and indexing, you can search directly on both local and SMB drives with the "live search" feature. Live search uses the same algorithm as the file scanning process. It ensures that the first match returned is always the fastest. In a Wi-Fi 2.4 test environment with an SMB folder of 1 million files, the results were typically 3x to 5x faster than the default Windows search tool.
5. Content-based search: In "live search" mode, when "search in file content" is enabled, DeepSearch will match the text content within the file. It supports the following file formats: PDF, Excel, Word, and plain text.
Access the link to download:
https://github.com/dohuyhoang93/DeepSearch/releases/tag/v1.2.0
Download DeepSearch.zip => Extract => Run DeepSearch.exe
Source: https://github.com/dohuyhoang93/DeepSearch