Local Extractive Text Summarizer
Local Extractive Text Summarizer: Scores sentences using TextRank graph algorithm based on lexical overlap and term significance.
About this local extractive text summarizer
Local Extractive Text Summarizer — browser-based utility.
How this tool works
Selects sentences using global repeated-token frequency divided by the square root of sentence token count, with a small first-sentence bonus. The current implementation is not TextRank or TF-IDF.
- Grapheme Cluster Segmentation: Segments text using Intl.Segmenter or UAX #29 boundaries to accurately treat multi-codepoint emojis and combining diacritics as single visual characters.
- Syntactic Chunking: Splits content into sentences using punctuation lookaheads while ignoring honorifics (e.g. 'Dr.', 'Inc.') and decimal numbers.
- Syllable & Entropy Calculation: Computes syllable counts per word and calculates Shannon lexical entropy ($H = -\sum p_i \log_2 p_i$) to measure vocabulary diversity.
- Index Synthesis & Metric Formatting: Produces target grade levels, estimated silent reading times (at 200-250 WPM), and speaking times (at 130-150 WPM).
Worked example
Scenario: A researcher generates a concise extractive summary from a long news article in the browser.
Sample input:
Processing: Scores sentences using TextRank graph algorithm based on lexical overlap and term significance.
Illustrative output:
Limits and verification
Handles unspaced scripts (such as Chinese, Japanese, and Thai) by switching to character/ideograph counting rather than space-delimited word tokens. Filters extreme text lengths gracefully up to browser memory limits.
Examples demonstrate an expected workflow; they do not prove every input or every branch of an external specification. Check important results with an independent source before using them for money, security, compliance, safety, or irreversible file changes.
Browser processing boundary
Tool input is processed by code running in the browser and is not intentionally sent to a CZOA processing API. The page can still request ordinary site assets, analytics, or advertising when those services are enabled. Browser extensions and managed-device software remain outside this tool's control.
Relevant references
These references govern or help explain the format, protocol, or calculation used here. Listing a reference does not claim certification or complete implementation of every optional feature.
- The Unicode Standard Version 15.1
- Standard Browser Web API / Algorithm Implementation (No single external RFC/ISO standard)
Content owner: CZOA Tools · Last reviewed: 2026-09-15 · Review methodology
How to use it
- Enter, paste, or select your input data into the Local Extractive Text Summarizer workspace controls.
- Review available parameter fields, units, formats, or options configured for your task.
- Click the action button or observe immediate live calculations rendered in your browser runtime.
- Inspect the resulting output and any diagnostic messages, then copy or download the result if needed.
Frequently asked questions
How does Local Text Summarizer select sentences?+
It splits on common sentence punctuation, scores sentences from repeated token frequency with a small first-sentence bonus, takes up to the requested limit, then restores source order.
What did the browser fixture verify?+
With three short sentences and the default limit, Cats sleep. Cats purr. Dogs bark. was returned in its original order.
Can the second input set summary length?+
Yes. It is converted to a number, truncated, limited between one and the available sentence count, and defaults to three when absent or unsuitable.
Does the result verify facts or understand document intent?+
No. This is local frequency scoring. It does not retrieve sources, resolve references, detect contradiction, preserve every caveat, or guarantee an abstractive summary.