Comments
>log in to commentHey DevHunt! I built rgx to bring ripgrep-style searching beyond plain text: PDFs, Office documents, ZIP archives, emails, media metadata, and GGUF/LLM files. Ripgrep-compatible CLI, zero external dependencies. Try it: rgx "search term" ./my_folder Paper + benchmark methodology: https://github.com/danalec/rgx-paper Speed comparisons require matching result counts—not just a faster stopwatch. Results vary by hardware and search pattern; rgx doesn’t beat ripgrep everywhere. For repeated searches, the optional prepared pack store delivered 1.55×–3.13× faster queries than ripgrep in published Linux tests, recovering its upfront build cost after roughly 21 queries on that setup. What formats or tricky files would you throw at it? Bug reports, feature requests, and honest feedback are welcome. Thanks for checking it out!
rgx is a ripgrep-compatible CLI that searches inside PDFs, Office docs, archives, emails, media metadata and LLM model files.
- for
- Developers who need fast, in-process full-text search of many binary formats via the command line.
- pricing
- open source
- license
- open source
- rgx-v0.1.1rgx v0.1.14 days ago
Key features
- Search inside many formats — Reads PDFs, Office documents, ZIP/RAR archives, emails, media metadata and GGUF/LLM files without external helpers.
- Ripgrep-compatible interface — Uses the same flags, output format and exit codes as ripgrep, so it drops into existing pipelines.
- Zero external dependencies — All adapters run in-process, avoiding spawning helper processes.
- Customizable builds — Cargo features let you include only the formats you need, reducing binary size.
- Filterable facts engine — Supports predicate-based filtering on file metadata such as mime type, author, EXIF GPS, etc.
- Optional facts index — Can build a JSON-Lines facts index for faster repeated queries on a stable corpus.
Use cases
- Search for a keyword inside a large collection of PDFs and Word documents
- Find passwords hidden inside compressed archives like .tar.gz or .zip
- Locate specific tensors or parameters inside LLM checkpoint files
- Filter emails by sender while grepping their content
- Run ad-hoc searches in CI pipelines that already use ripgrep
rgx FAQ
How does rgx differ from ripgrep?+
rgx adds in-process adapters that can unpack and search inside binary containers, while ripgrep treats those files as opaque bytes.
Can I limit the binary size?+
Yes, you can disable default features and enable only the formats you need via Cargo feature flags.
Is there an index for faster repeated searches?+
rgx can build a JSON-Lines facts index with --facts-build and query it later with --facts-index.
What output does rgx produce?+
It uses the same path:line:text format as ripgrep, making it compatible with existing tools.
How do I install rgx on macOS?+
Install via Homebrew with `brew install rgx` or use the provided formula in the repository.
Summarized by DevHunt from github.com · Oct 31, 2026. Details may change; check the official site.
