How to use Named Entity Recognition Tool
Paste text into Input text or use Upload for a .txt, .csv, .json, .log, or .md file. Named Entity Recognition Tool groups matches for people, organizations, locations, dates, money, percentages, and times. Each row includes the category, occurrence count, and source positions, and Output text contains the same rows as tab-separated text.
- Paste text into Input text, or select Upload and choose a
.txt,.csv,.json,.log, or.mdfile. - Keep All categories selected, or choose Person, Organization, Location, Date, Money, Percent, and Time individually.
- Read Detected entities and Output text. The table shows Entity, Category, Count, and Source Position.
- Select Copy or Download to take the tab-separated entity list into another document.
What Named Entity Recognition Tool extracts
Person matching accepts a title followed by a capitalized name, or a two- or three-word capitalized name. Organization matching checks names with suffixes such as Inc. and University, along with listed acronyms. Location matching checks listed countries, US states, cities, and names ending in terms such as River, Lake, or County.
The other categories use formatted patterns. Dates include 2024-03-15 and March 15, 2024. Money includes $1 billion and 250,000 USD; percentages include 12%; times include 9:00 AM.
How to read the extracted entity list
This input produces one row for each category-and-text match. The first spelling found is shown, with its count and character position:
Tim Cook met Apple Inc. in Austin on March 15, 2024 at 9:00 AM and noted 12% growth.12% Percent 1 73
9:00 AM Time 1 55
Apple Inc. Organization 1 13
Austin Location 1 27
March 15, 2024 Date 1 37
Tim Cook Person 1 0Repeated matches are grouped without regard to letter case, while the category stays part of the group. Count records every occurrence. Source Position lists the first five character indexes, followed by +N more when additional positions exist.
Where Named Entity Recognition Tool can miss
Regular expressions, fixed lists, capitalization rules, titles, organization suffixes, acronyms, and stopwords determine what the tool can recognize. Unsupported names, unusual wording, lowercase names, and ambiguous capitalization can be missed or classified differently. When candidate spans overlap, the priority order keeps one match: Date, Money, Percent, Time, Organization, Location, then Person.
Select Reset categories to turn every category back on after narrowing the selection. With empty input or no checked categories, the result count is zero and Output text is blank.
Examples of text to check
An editor can scan a draft for people, organizations, and locations before checking names against a source list. A researcher can inspect dates, money, and percentages in a report, then download the rows for review.
A developer can paste a log excerpt and check for times or locations. A records worker can upload a Markdown or CSV file, clear unrelated categories, and copy the remaining rows.
Named Entity Recognition Tool compared with other extractors
Use Text Parsing & Data Extraction when you are choosing among text-specific jobs. Use Date Extractor & Format Harmonizer when dates are the target and you need List, Harmonize, Table, CSV, or JSON output. Choose Keyword List Extractor for exact keywords you provide, with occurrence counts, line numbers, and context. Choose Custom Regex Extractor when your pattern does not fit these fixed entity rules.
Frequently asked questions
What categories can Named Entity Recognition Tool find?
It finds Person, Organization, Location, Date, Money, Percent, and Time matches. Each category can be selected separately.
How are entity counts and source positions shown?
Each unique category-and-text match gets one row with its count and character indexes. The output shows up to five indexes, then reports extra positions as +N more.
Can I extract only dates or organizations?
Yes, clear All categories and select only Date or Organization. The results update with the categories that remain checked.
Does Named Entity Recognition Tool send my text to a server?
No network request is made for extraction. The browser reads pasted text and uploaded files with JavaScript and browser APIs.
Why was a name or place not detected?
The matcher relies on fixed patterns and lists, so unsupported names, wording, capitalization, or overlapping matches can be missed. Check the source text and category selection before using the list.