A high-performance credential extractor that parses massive url:user:pass files and extracts matching entries by domain — boundary-aware matching, multi-threaded I/O, zero-copy memory maps.

---

Features

Category	Detail
Domain matching	Boundary-aware + case-insensitive — subdomains, URLs, emails; rejects false positives
Performance	Memory-mapped I/O, rayon parallelism, SIMD-accelerated byte search, live per-chunk progress
I/O	Single-file, multi-file batch (-a), recursive directory walk (-r)
Output control	Deduplication, append mode (-A), match limit (-M), quiet mode (-q)
Safety	Refuses to write when output collides with an input path
Formats	Configurable ASCII divider — :, |, ;, or any printable ASCII separator
UX	Styled CLI with live progress bar, full interactive TUI, graceful cancel
Platform	Linux, macOS, Windows — pre-built binaries on every release

Quick Start

# Pre-built binary (recommended)
curl -LO https://github.com/sinescode/ulpExtractor/releases/latest/download/ulpExtractor-linux-x86_64.tar.gz
tar xzf ulpExtractor-linux-x86_64.tar.gz
./ulpExtractor

# Or build from source
git clone https://github.com/sinescode/ulpExtractor.git && cd ulpExtractor
cargo build --release
./target/release/ulpExtractor

Usage

Single File

ulpExtractor -d netflix.com -i combo.txt
ulpExtractor -d netflix.com -i combo.txt -o results.txt
ulpExtractor -d netflix.com -i huge_dump.txt -M 100       # limit to 100 matches
ulpExtractor -d netflix.com -i combo.txt -q               # no progress bar

Batch Scan

ulpExtractor -d netflix.com -a                             # all .txt in current dir
ulpExtractor -d netflix.com -a -x txt,csv,json              # specific extensions
ulpExtractor -d netflix.com -a -r                           # recursive
ulpExtractor -d netflix.com -a --dir ./data -t 8            # custom dir, 8 threads
ulpExtractor -d netflix.com -i extra.txt -A                 # append to existing output

Interactive TUI

ulpExtractor                 # empty form
ulpExtractor --tui -d netflix.com -i dump.txt   # pre-filled form

Keyboard

Key	Action
Tab / ↑ ↓	Move between fields
← → / Space	Toggle mode, checkboxes, advanced
Enter	Run extraction (or toggle focused control)
?	Help overlay
Esc / q	Cancel extraction or quit
r	Re-run last config (results screen)
Ctrl+U	Clear focused text field
Ctrl+W	Delete previous word

Options

Flag	Description	Default
-d, --domain	Domain to extract credentials for (normalized, case-insensitive)	required
-i, --input	Input file path	—
-a, --all	Scan all files matching extensions in a directory	—
-r, --recursive	Walk directories recursively (with -a)	—
-x, --extensions	File extensions to include (comma-separated)	txt
--dir	Target directory for -a	.
-o, --output	Output file path	output.txt
-A, --append	Append to output instead of overwriting	—
-t, --threads	Number of worker threads (capped at 64)	4
-D, --divider	Field separator character (printable ASCII)	:
-M, --max-matches	Stop after N unique matches	unlimited
-q, --quiet	Suppress the progress bar	—
--tui	Open the interactive TUI (optionally pre-filled)	—

Input Format

<url_or_domain><divider><user><divider><password>

The domain may appear as a bare host, subdomain, inside a URL path, or in an email:

netflix.com:john:secret123
www.netflix.com:user@mail.com:pass456
https://platform.deepseek.com/login:admin:pass789
user@example.com:somepass

Matching is boundary-aware and case-insensitive — deepseek.com matches platform.DeepSeek.COM but not mydeepseek.com. Output is user<divider>password, one per line, deduplicated.

Domain input is normalized: schemes (https://), paths, userinfo, and trailing ports are stripped before matching.

Safety

• Output path must not collide with any input file (prevents truncating dumps).
• The TUI asks before overwriting an existing non-append output file.
• Cancel (Esc / Ctrl-C) flushes partial unique matches already found.

Build

Requires Rust 1.70+ (rustup).

cargo build --release
# → target/release/ulpExtractor