Sentence Segmenter

Overview

Splits your text into sentences with Unicode-standard, locale-aware boundary rules, so decimals like 3.79, multi-dot abbreviations like U.S.A., and ellipses stay inside one sentence instead of splitting it. You get the sentences in order, each with its index and, unless you turn offsets off, the exact character range it occupies in your text. Accepts file input via file_ref — no need to extract file content first.

Sentence boundaries are tricky — abbreviations, decimals, and ellipses fool naive split-on-period rules. We use ICU4J so you get linguistically correct boundaries without LLM guessing.

When to use it

Scope

Non-goals

• Does not detect language — provide language hint or use text.detect_language first
• Does not summarize or paraphrase — output is verbatim text segments
• Does not perform word tokenization — use a dedicated tokenizer for that

Properties

Capabilities

Cost & billing

Flat rate per invocation

File handling

File limits

Constraints this skill applies to your input and generated files. The “Limited by” column shows the static source of each cap; none of these can be raised by upgrading your plan.

How we handle your data

Input schema

Output schema

API contract

curl -sS -X POST "$BASE/v1/execute" -H "Authorization: Bearer $KEY" -H 'Content-Type: application/json' -d '{"skillKey":"text.segment_sentences","input":{"text":"Dr. Smith arrived. The meeting started.","language":"en"}}'

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

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

const { output } = await sq.execute("text.segment_sentences", {"text":"Revenue grew 12.5% last quarter. The board approved the plan... Next review is in June."});

Errors

Other limits

Limitations

• Does not detect language — provide language hint or use text.detect_language first
• Does not summarize or paraphrase — output is verbatim text segments
• Does not perform word tokenization — use a dedicated tokenizer for that
• Paragraph-level splitting (use line-break heuristics instead)
• Word tokenization or morphological analysis
• Summarization or content reduction

How to call

POST /v1/execute (requires authentication)

{
  "skillKey": "text.segment_sentences",
  "input": {
    "text": "Revenue grew 12.5% last quarter. The board approved the plan... Next review is in June."
  }
}

Examples

{
  "text": "Revenue grew 12.5% last quarter. The board approved the plan... Next review is in June."
}

{
  "count": 3,
  "sentences": [
    {
      "text": "Revenue grew 12.5% last quarter.",
      "startOffset": 0,
      "endOffset": 33,
      "index": 0
    },
    {
      "text": "The board approved the plan...",
      "startOffset": 33,
      "endOffset": 64,
      "index": 1
    },
    {
      "text": "Next review is in June.",
      "startOffset": 64,
      "endOffset": 87,
      "index": 2
    }
  ]
}

{
  "includeOffsets": false,
  "text": "The U.S.A. spans 3.79 million square miles. It has 50 states."
}

{
  "count": 2,
  "sentences": [
    {
      "index": 0,
      "text": "The U.S.A. spans 3.79 million square miles."
    },
    {
      "index": 1,
      "text": "It has 50 states."
    }
  ]
}

{
  "file_ref": "squillify://files/f-abc123"
}

{
  "count": 48,
  "sentences": [
    {
      "text": "First sentence.",
      "startOffset": 0,
      "endOffset": 16,
      "index": 0
    }
  ]
}

Recommended use cases

• Split text into sentences for readability scoring or analysis
• NLP preprocessing requiring sentence-level boundaries
• Chunk long documents into sentence-aligned segments for embedding

Avoid for

• Paragraph-level splitting (use line-break heuristics instead)
• Word tokenization or morphological analysis
• Summarization or content reduction

Related skills

Composable with

Safety & privacy

• Runs without calling any external service.
• Your payloads are never stored permanently.
• Telemetry captures metadata only — never your payload contents.
• We never train models on your payloads.
• No human reviews your payloads.

Tags