Extract Summary Sentences

text.extract_summary_sentences

ProductionEngine wrapper (Class A)Deterministicstrictmoderatev1.0.0

No network

Overview

You send prose up to two mebibytes and we return a subset of its own sentences, never invented wording. ICU sentence and word boundaries score each sentence by in-document term frequency after stopword removal. maxSentences defaults to 5 and tops out at 50. Optional maxCharacters refuses to add a sentence that would overflow the budget rather than truncating it. Spans are original character offsets so slicing the source reconstructs each selected sentence. This does not paraphrase, translate, or parse HTML. A successful call costs 2 execution credits.

Agents need a published extractive subset of existing sentences, not a paraphraser.

Reliability

Sentence and word boundaries follow the pinned ICU 76.1 locale data. TEXT_SUMMARY_LIMIT_EXCEEDED — trigger: text longer than 2097152 characters, more than 50 maxSentences, or more than 2000 stopwords; retryable: no; next: split the source or lower the cap. TEXT_SUMMARY_INVALID — trigger: a blank locale or a stopword outside 1-64 characters; retryable: no; next: send a BCP-47 tag and bounded stopwords.

Context footprint

Returned sentences are original substrings. A successful call costs 2 execution credits.

When to use it

When a caller already holds prose and needs a bounded extractive highlight with original spans.

When the task is abstractive rewriting, HTML parsing, or factuality checking.

Scope

Note

Selects existing sentences as an extractive summary using ICU boundaries and in-document term frequency.

Non-goals

Does not paraphrase or invent wording
Does not parse HTML or documents
Does not assess factuality

Properties

IdempotentRetry safeStreaming supportedFile input supportedRequires external service

Capabilities

analysistext-processing

Cost & billing

Flat rate per invocation

MetercallsBillable onsuccessCost per call2 credits

File handling

Accepts file_refProduces file_refMax input2 MB

How we handle your data

Sensitive output possible

What this means

This skill may receive sensitive content and may return submitted, transformed, extracted, or reformatted content.

Squillify guarantees

Runs inside Squillify
This skill runs inside Squillify.
Payloads are not sent to an external service by this skill.
Squillify does not permanently store submitted payloads.
Telemetry records metadata only, not submitted content.
Submitted payloads are not used for training.
Submitted payloads are not reviewed by humans.
This skill does not target sensitive data.
The response may contain submitted or transformed content.

Your responsibility

Squillify does not determine whether submitted content is sensitive. Your application is responsible for deciding what data it is allowed to submit. Your application is responsible for securing the response.

Read the Microskill Data Handling Policy

Input schema

Field	Type	Required	Description
text	string	Yes	Source prose, 0 through 2097153 characters; runtime accepts at most 2097152.
locale	string	No	BCP-47 tag selecting ICU sentence and word rules; omit it for en.
maxSentences	integer	No	How many sentences to keep, 1 through 51; omit it for 5. Runtime accepts at most 50.
maxCharacters	integer	No	Concatenated-summary character budget, 1 through 2097153. Omit it for no extra cap.
preserveOrder	boolean	No	True emits selected sentences in document order; omit it for true.
stopwords	array	No	Caller stopwords, 0 through 2001 strings of 1 through 64 characters that replace the built-in English list. Omit it to use the built-in list.

Output schema

Field	Type	Required	Description
sentences	array	Yes	Selected sentences in the applied order, 0 through 50 closed {text, start, end, score, index} objects.
summary	string	Yes	Selected sentence texts joined with a single space, 0 through 2097152 characters.
locale	string	Yes	BCP-47 locale tag that was actually applied to segmentation and scoring.
maxSentences	integer	Yes	Sentence cap that was actually applied to this call.
preserveOrder	boolean	Yes	Whether selected sentences were emitted in document order.
candidateCount	integer	Yes	How many non-blank sentences ICU found in the source.
selectedCount	integer	Yes	How many sentences were kept after ranking and optional character budget.
engine	string	Yes	Pinned ICU engine coordinate used for this answer.

API contract

POST /v1/execute

curl

curl -sS -X POST "$BASE/v1/execute" -H "Authorization: Bearer $KEY" -H 'Content-Type: application/json' -d '{"skillKey":"text.extract_summary_sentences","input":{"text":"Cats chase mice. Dogs chase cats. Mice hide from cats."}}'

Integrations

TypeScript SDK

import { Squillify } from "@squillify/sdk";

const sq = new Squillify({ apiKey: process.env.SQUILLIFY_API_KEY });

const { output } = await sq.execute("text.extract_summary_sentences", {"text":"Cats chase mice. Dogs chase cats. Mice hide from cats."});

Errors

Code	HTTP	Retryable	Description
VALIDATION_ERROR	400	No	The request payload failed schema validation.
UNAUTHORIZED	401	No	Missing or invalid API key.
SKILL_NOT_ACCESSIBLE	403	No	The API key is not granted access to this skill.
SKILL_NOT_AVAILABLE_FOR_PLAN	403	No	This skill is not included in your current plan.
PLAN_ELIGIBILITY_UNAVAILABLE	503	Yes	Plan eligibility could not be decided. Retry after backoff.
SKILL_NOT_FOUND	404	No	The requested skill or version does not exist.
TENANT_RATE_LIMITED	429	Yes	Request rate limit exceeded. Retry after backoff.
QUOTA_EXCEEDED	429	Yes	Plan quota exhausted. Retryable once the quota period resets.
EXECUTION_TIMEOUT	504	Yes	Execution exceeded the allowed time budget.
TEXT_SUMMARY_LIMIT_EXCEEDED	429	No	A size, quota, rate, or time limit was exceeded.
TEXT_SUMMARY_INVALID	400	No	The request payload failed this skill's input validation.

Other limits

Limit	Value
Recommended inline bytes	32 KiB

Limitations

Does not paraphrase or invent wording
Does not parse HTML or documents
Does not assess factuality
Abstractive summarization
HTML or PDF extraction

How to call

POST /v1/execute (requires authentication)

Request bodyjson

{
  "skillKey": "text.extract_summary_sentences",
  "input": {
    "text": "Cats chase mice. Dogs chase cats. Mice hide from cats."
  }
}

Examples

Three short sentences keep the two that mention cats most

Inputjson

{
  "text": "Cats chase mice. Dogs chase cats. Mice hide from cats."
}

Expected outputjson

{
  "locale": "en",
  "maxSentences": 5,
  "candidateCount": 3,
  "sentences": [
    {
      "text": "Cats chase mice. ",
      "start": 0,
      "end": 17,
      "score": 2.333333,
      "index": 0
    },
    {
      "text": "Dogs chase cats. ",
      "start": 17,
      "end": 34,
      "score": 2,
      "index": 1
    },
    {
      "text": "Mice hide from cats.",
      "start": 34,
      "end": 54,
      "score": 2,
      "index": 2
    }
  ],
  "preserveOrder": true,
  "engine": "com.ibm.icu:icu4j",
  "selectedCount": 3,
  "summary": "Cats chase mice.  Dogs chase cats.  Mice hide from cats."
}

Tip

Frequency scoring prefers sentences that repeat content words already in the document.

Whitespace-only prose yields no candidate sentences