Text Processing

Direct text processing endpoints allow you to translate, revise, and rewrite text synchronously without creating background jobs. These endpoints return results immediately, making them ideal for real-time applications. Streaming variants are available for translate and rewrite, delivering results via Server-Sent Events.

POST/v1/text/translate

Translate text

This endpoint allows you to translate text directly using AI. It accepts a list of source texts and returns a corresponding list of translations.

Required attributes

Name
source_text
Type
array of strings
Description
List of texts to translate.
Name
source_language
Type
string
Description
The source language code (e.g., "en").
Name
target_language
Type
string
Description
The target language code (e.g., "de").

Optional attributes

Name
domain
Type
string
Description
Content domain for the translation (e.g., "technical", "legal", "marketing").
Name
provider
Type
string
Description
The AI provider to use: openai, anthropic, mistral, gemini, lara, widn (default: openai).
Name
context_window
Type
integer
Description
Processing mode (1=full_document, 2=bucketing, 3=segment_per_segment, default: 3).
Name
glossaries
Type
array of integers
Description
List of glossary IDs to use for terminology consistency.
Name
context_stores
Type
array of integers
Description
Context store IDs for styleguide content.
Name
instructions
Type
string
Description
Additional instructions for the translation.
Name
extra_instructions
Type
array of strings
Description
Additional instructions passed to the translation provider.

Request

POST

/v1/text/translate

curl -X POST https://api.elanlanguages.ai/v1/text/translate \
  -H "Authorization: Bearer {token}" \
  -H "X-Org-Id: {orgId}" \
  -H "Content-Type: application/json" \
  -d '{
    "source_text": ["Welcome to our product documentation."],
    "source_language": "en",
    "target_language": "de",
    "domain": "technical",
    "glossaries": [1]
  }'

Response

{
  "status": "success",
  "source_text": ["Welcome to our product documentation."],
  "translated_text": ["Willkommen in unserer Produktdokumentation."],
  "source_language": "en",
  "target_language": "de",
  "provider": "openai",
  "domain": "technical",
  "context_window": 3,
  "processing_mode": "direct"
}

POST/v1/text/translate/stream

Stream translate text

This endpoint streams translated text via Server-Sent Events. It accepts the same request body as the synchronous translate endpoint but streams the translation of the first text in the source_text list.

The stream emits three event types:

delta — A chunk of translated text: {"text": "..."}
done — The final result with metadata: {"text": "...", "source_language": "...", ...}
error — An error occurred: {"error": "...", "status": 402|400|500}

Required attributes

Name
source_text
Type
array of strings
Description
List of texts to translate. The stream translates the first item.
Name
source_language
Type
string
Description
The source language code (e.g., "en").
Name
target_language
Type
string
Description
The target language code (e.g., "de").

Optional attributes

Name
domain
Type
string
Description
Content domain for the translation (e.g., "technical", "legal", "marketing").
Name
provider
Type
string
Description
The AI provider to use: openai, anthropic, mistral, gemini, lara, widn (default: openai).
Name
context_window
Type
integer
Description
Processing mode (1=full_document, 2=bucketing, 3=segment_per_segment, default: 3).
Name
glossaries
Type
array of integers
Description
List of glossary IDs to use for terminology consistency.
Name
context_stores
Type
array of integers
Description
Context store IDs for styleguide content.
Name
instructions
Type
string
Description
Additional instructions for the translation.
Name
extra_instructions
Type
array of strings
Description
Additional instructions passed to the translation provider.

Request

POST

/v1/text/translate/stream

curl -N -X POST https://api.elanlanguages.ai/v1/text/translate/stream \
  -H "Authorization: Bearer {token}" \
  -H "X-Org-Id: {orgId}" \
  -H "Content-Type: application/json" \
  -d '{
    "source_text": ["Welcome to our product documentation."],
    "source_language": "en",
    "target_language": "de"
  }'

Response (SSE)

event: delta
data: {"text": "Willkommen"}

event: delta
data: {"text": " in unserer"}

event: delta
data: {"text": " Produktdokumentation."}

event: done
data: {"text": "Willkommen in unserer Produktdokumentation.", "source_language": "en", "target_language": "de"}

POST/v1/text/revise

Revise text

This endpoint allows you to revise translated text directly using AI post-editing. It provides synchronous text revision without TMS integration, using the same post-editing adapters and processing logic as the AIPE handler.

Required attributes

Name
source_text
Type
string
Description
The original source text.
Name
target_text
Type
string
Description
The existing translation to revise.
Name
source_language
Type
string
Description
The source language code (e.g., "en").
Name
target_language
Type
string
Description
The target language code (e.g., "de").

Optional attributes

Name
domain
Type
string
Description
Content domain for the revision (e.g., "technical", "legal", "marketing").
Name
provider
Type
string
Description
The AI provider to use: openai, anthropic, mistral, gemini (default: openai).
Name
context_window
Type
integer
Description
Processing mode (1=full_document, 2=bucketing, 3=segment_per_segment, default: 3).
Name
glossaries
Type
array of integers
Description
List of glossary IDs to use for terminology consistency.
Name
context_stores
Type
array of integers
Description
Context store IDs for styleguide content.
Name
instructions
Type
string
Description
Additional instructions for the revision.
Name
styleguide
Type
string
Description
Direct styleguide text to follow for the revision. Used if context_stores is not provided.

Request

POST

/v1/text/revise

curl -X POST https://api.elanlanguages.ai/v1/text/revise \
  -H "Authorization: Bearer {token}" \
  -H "X-Org-Id: {orgId}" \
  -H "Content-Type: application/json" \
  -d '{
    "source_text": "Welcome to our product documentation.",
    "target_text": "Wilkommen zu unsere Produktdokumentation.",
    "source_language": "en",
    "target_language": "de",
    "domain": "technical"
  }'

Response

{
  "status": "success",
  "source_text": "Welcome to our product documentation.",
  "original_target_text": "Wilkommen zu unsere Produktdokumentation.",
  "revised_text": "Willkommen in unserer Produktdokumentation.",
  "source_language": "en",
  "target_language": "de",
  "provider": "openai",
  "domain": "technical",
  "context_window": 3,
  "processing_mode": "direct"
}

POST/v1/text/rewrite

Rewrite text

This endpoint allows you to rewrite text using AI. You specify a rewrite type to control how the text is transformed — for example, simplifying complex language, summarizing long passages, or elaborating on brief content.

Required attributes

Name
source_text
Type
string
Description
The text to rewrite.
Name
source_language
Type
string
Description
The language code of the text (e.g., "en").
Name
type
Type
string
Description
The rewrite type. One of: simplify, summarize, elaborate, continue, align.

Optional attributes

Name
domain
Type
string
Description
Content domain for the rewrite (e.g., "technical", "legal", "marketing").
Name
glossaries
Type
array of integers
Description
List of glossary IDs to use for terminology consistency.
Name
context_stores
Type
array of integers
Description
Context store IDs for styleguide content.
Name
instructions
Type
string
Description
Additional rewriting instructions.
Name
template
Type
integer
Description
Explicit prompt template ID override.

Request

POST

/v1/text/rewrite

curl -X POST https://api.elanlanguages.ai/v1/text/rewrite \
  -H "Authorization: Bearer {token}" \
  -H "X-Org-Id: {orgId}" \
  -H "Content-Type: application/json" \
  -d '{
    "source_text": "The implementation of the aforementioned functionality necessitates the utilization of advanced computational paradigms.",
    "type": "simplify",
    "source_language": "en"
  }'

Response

{
  "status": "success",
  "source_text": "The implementation of the aforementioned functionality necessitates the utilization of advanced computational paradigms.",
  "rewritten_text": "Building this feature requires advanced computing techniques.",
  "source_language": "en",
  "domain": null,
  "type": "simplify",
  "template": null
}

POST/v1/text/rewrite/stream

Stream rewrite text

This endpoint streams rewritten text via Server-Sent Events. It accepts the same request body as the synchronous rewrite endpoint but delivers results incrementally as they are generated.

The stream emits three event types:

delta — A chunk of rewritten text: {"text": "..."}
done — The final result with metadata: {"text": "...", "source_language": "...", ...}
error — An error occurred: {"error": "...", "status": 402|400|500}

Required attributes

Name
source_text
Type
string
Description
The text to rewrite.
Name
source_language
Type
string
Description
The language code of the text (e.g., "en").
Name
type
Type
string
Description
The rewrite type. One of: simplify, summarize, elaborate, continue, align.

Optional attributes

Name
domain
Type
string
Description
Content domain for the rewrite (e.g., "technical", "legal", "marketing").
Name
glossaries
Type
array of integers
Description
List of glossary IDs to use for terminology consistency.
Name
context_stores
Type
array of integers
Description
Context store IDs for styleguide content.
Name
instructions
Type
string
Description
Additional rewriting instructions.
Name
template
Type
integer
Description
Explicit prompt template ID override.

Request

POST

/v1/text/rewrite/stream

curl -N -X POST https://api.elanlanguages.ai/v1/text/rewrite/stream \
  -H "Authorization: Bearer {token}" \
  -H "X-Org-Id: {orgId}" \
  -H "Content-Type: application/json" \
  -d '{
    "source_text": "The implementation of the aforementioned functionality necessitates the utilization of advanced computational paradigms.",
    "type": "simplify",
    "source_language": "en"
  }'

Response (SSE)

event: delta
data: {"text": "Building this"}

event: delta
data: {"text": " feature requires"}

event: delta
data: {"text": " advanced computing techniques."}

event: done
data: {"text": "Building this feature requires advanced computing techniques.", "source_language": "en", "type": "simplify"}