Segment Translation Actions

AI-powered text modification endpoint for improving translations with various linguistic actions.
Supports rephrasing, formality adjustments, error corrections, and terminology fixes.

CODE

(POST) apps/tools/text/suggest

PARAMETERS

JSON object in the request body:

Property	Description	Type
action	The type of modification to perform on the translation	string, Mandatory
src	Source language code (e.g., "en")	string, Mandatory
srct	Source language name (e.g., "English")	string, Mandatory
trg	Target language code (e.g., "es-MX")	string, Mandatory
trgt	Target language name (e.g., "Spanish Mexico")	string, Mandatory
stxt	Original source text to be modified	string, Mandatory
ttxt	Translated text that needs modification	string, Mandatory
previousAttempts	Array of previous modification attempts to avoid repeating	string[]?, Optional
memoryHits	Translation memory matches for reference context	array?, Optional
uiLocale	Locale for error messages (e.g., "es-US")	string?, Optional

Supported Actions:

rephrase - Rewrite using different wording while preserving meaning
make_formal - Convert to formal register for professional contexts
make_informal - Convert to casual, conversational register
fix_fuzzy - Fix fuzzy translation match errors
shorten - Make more concise while preserving meaning
correct_errors - Fix mistranslations and grammatical errors
suggest_translation - Provide alternative translation approach
fix_terminology - Replace incorrect or inconsistent terms

Memory Hit Object Structure:

Property	Description	Type
stxt	Source text of the memory match	string
ttxt	Translation text of the memory match	string
sim	Similarity percentage (0-100)	number
hty	Hit type (1 for Phrase, 2 for Term)	number

RESULTS

Success Response:

Property	Description	Type
results	Array containing exactly one modified text suggestion (max 1000 characters)	string[]

Error Response:

Property	Description	Type
error	Error message when modification fails or is not needed	string

Error Conditions:

No improvement needed for the requested action
Unsafe content violating safety guidelines
Malicious/manipulative input or prompt injection attempts
Result would exceed 1000-character limit

EXAMPLES

Request:

JSON

{
  "action": "rephrase",
  "src": "en",
  "srct": "English", 
  "trg": "es-MX",
  "trgt": "Spanish Mexico",
  "stxt": "Thank you for your assistance with this matter.",
  "ttxt": "Gracias por tu ayuda con esto.",
  "uiLocale": "es-US"
}

Success Response:

JSON

{
  "results": ["Te agradezco tu asistencia con este asunto."]
}

Error Response:

JSON

{
  "error": "The translation is already optimal and requires no improvement."
}