Data extraction
Regex Extractor
Turn a regular-expression pattern into copyable matches without running arbitrary JavaScript.
Processed locally. Your input and output are never uploaded.
Tool overview
How guarded matching and capture output work
Regex Extractor compiles a length-limited JavaScript regular expression inside a local Web Worker. It shows whole-match previews, code-unit positions, capture values, and named groups. Suspicious repetition is rejected, zero-length matches advance safely, and long-running evaluation is terminated.
Capture identifiers with a known pattern
- Extract identifiers from log lines
- Collect one capture group from repeated records
- Inspect named groups while designing a pattern
- Export match positions for a text-processing workflow
Run a guarded regular-expression match
- Paste the source text.
- Enter a regular expression and supported flags.
- Choose whole, selected, all, or named capture output.
- Review the preview table, match positions, and counts.
- Copy or download the extracted values.
Flags, captures, named groups, and separators
- Global matching is always enabled; other supported JavaScript flags can be entered explicitly.
- Selected capture mode uses a one-based group number.
- All capture groups are tab-separated per match, while named mode labels each value.
- Evaluation is limited to a 200-character pattern, 1 MB input, 100,000 matches, and a worker timeout.
Export one selected capture group
Before
order-17, order-42
After
id=17 id=42
Pattern safety, timeout, and input limits
- Positions are zero-based UTF-16 code-unit offsets, matching JavaScript regular-expression behavior.
- An empty match is represented safely and the scanner advances to avoid an infinite loop.
- Nested-repetition checks are conservative and cannot predict every expensive pattern.
- A timed-out worker is terminated and no partial output is returned.
Browse all data extraction tools →
Questions
Regex Extractor FAQ
Can the pattern execute JavaScript?
No. The tool accepts only regular-expression source and flags, then returns matched text.
What happens when matching takes too long?
The worker is stopped and the tool reports that the regular expression took too long.