Your privacy choices

We use essential storage to operate Lexel Translate. With your permission, we also use analytics, advertising, and marketing technologies. Read our Cookie Policy.

Translate

Translate one logical text payload synchronously with glossary selection, structured-content handling, and placeholder protection.

Endpoint

POST/api/v1/translate

Native translate endpoint.

Request example

POST/api/v1/translate

curl -X POST "https://translate.lexel.ai/api/v1/translate" \
  -H "Authorization: Bearer lxl_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
  "request_id": "req_translate_demo_001",
  "src": "en-Latn",
  "tgt": "fr-Latn",
  "content_mode": "plain",
  "preserve_placeholders": true,
  "do_not_translate": [
    "{name}"
  ],
  "metadata": {
    "order_id": "ord_123"
  },
  "input": {
    "items": [
      {
        "txt": "Open a support ticket for {name}."
      }
    ]
  }
}'

Response example

json

{
  "ok": true,
  "request_id": "req_translate_demo_001",
  "operation": "translate",
  "source_language_tag": "en-Latn",
  "source_language_name": "English",
  "detection_confidence": 0.99,
  "inference_latency": 142,
  "output": {
    "kind": "single",
    "single": {
      "translation": "Ouvrez un ticket d'assistance pour {name}.",
      "src": "en-Latn",
      "tgt": "fr-Latn"
    }
  },
  "usage": {
    "input_characters": 33,
    "output_characters": 42
  },
  "warnings": [],
  "quota": {
    "creditType": "premium",
    "membership": "premium",
    "monthlyCharacterQuota": 5000000,
    "monthlyCharacterUsage": 120033,
    "monthlyBillableCharacterUsage": 0,
    "perRequestCharacterLimit": null
  },
  "trial": {
    "isAuthenticated": true,
    "limit": null,
    "used": null,
    "remaining": null,
    "startedAt": null,
    "windowMs": null
  }
}

Request fields

Use one item in input.items for synchronous single translation. Root-level config keeps the public parser consistent with batch and async routes.

Field	Type	Required	Description
request_id	string	Yes	Caller-provided trace ID returned in responses and logs.
src	language tag \| ""	No	Root source language for the request. Use an empty string to auto-detect.
tgt	language tag	Yes	Root target language for the request.
content_mode	auto \| plain \| markdown \| html \| json \| xml \| yaml \| csv	No	Structured content is preprocessed before translation.
split_input	0 \| 1 \| 2	No	0 off, 1 balanced (default), 2 aggressive. Splitting can reduce latency with generally minimal impact on translation quality.
ps_context	string	No	Optional preceding-sentence context for short or ambiguous text.
metadata	object	No	Caller-defined JSON metadata echoed in logs and responses when present.
glossary_id	string	No	Stored glossary selection for this request.
glossary_version_id	string	No	Optional immutable stored glossary version for reproducible sync behavior.
glossary_group_ids	string[]	No	Selected enabled glossary groups. Combined with item selectors as a union.
glossary_item_ids	string[]	No	Selected enabled glossary items. Combined with group selectors as a union.
glossary.entries	array	No	Inline glossary mappings when you want request-scoped terminology instead of a stored glossary.
glossary.do_not_translate	string[]	No	Inline protected terms carried with the inline glossary.
do_not_translate	string[]	No	Top-level protected terms such as placeholders or brand names.
preserve_placeholders	boolean	No	Disabled by default. When true, the app detects placeholder-like patterns and adds them to protection rules without rewriting the source text.
strict_placeholder_preservation	boolean	No	Only applies when preserve_placeholders is true. Tightens enforcement when protected values are not preserved.
input.items	array(length = 1)	Yes	Native public single-translate requests use one item in the items array.
input.items[0].txt	string	Yes	Source text to translate.

Response fields

Field	Type	Required	Description
ok	boolean	Yes	True for successful translate responses.
request_id	string	Yes	Echoes the caller request ID or a generated request ID.
operation	"translate"	Yes	Response operation discriminator.
output.kind	"single"	Yes	Single-output discriminator.
output.single.translation	string	Yes	Canonical translated text field.
output.single.src	string	No	Canonical source language tag for the returned translation.
output.single.tgt	string	No	Canonical target language tag for the returned translation.
source_language_tag	string	No	Canonical supplied or detected source language tag.
source_language_name	string	No	Human-readable source language name.
detection_confidence	number \| null	No	Provider-backed detection confidence when available.
inference_latency	number \| null	No	Provider-reported inference latency in milliseconds when available.
usage.input_characters	number	Yes	Characters counted from the request input.
usage.output_characters	number	Yes	Characters in translated output.
warnings	array	Yes	Non-fatal warnings such as preservation or confidence notes.
quota	object \| null	No	Quota state returned for authenticated actors when available.
trial	object \| null	No	Runtime quota-window state returned in the current response envelope.

Errors

json

{
  "ok": false,
  "error": "Premium characters exhausted.",
  "status": 429,
  "request_id": "req_translate_demo_001",
  "quota": {
    "creditType": "premium",
    "membership": "premium",
    "monthlyCharacterQuota": 5000000,
    "monthlyCharacterUsage": 5000000,
    "monthlyBillableCharacterUsage": 0,
    "perRequestCharacterLimit": null
  }
}

Behavior

Structured content modes such as json and csv are preprocessed before translation. The provider sees extracted text segments, not your original structure.
Placeholder detection never rewrites the input text. When enabled, it only detects placeholder-like values and adds them to protected-term handling.

Next steps

See Synchronous batch for multiple items in one synchronous request.
See Glossaries for stored terminology selection.

Translate

Translate one logical text payload synchronously with glossary selection, structured-content handling, and placeholder protection.

Endpoint

POST/api/v1/translate

Native translate endpoint.

Request example

POST/api/v1/translate

curl -X POST "https://translate.lexel.ai/api/v1/translate" \
  -H "Authorization: Bearer lxl_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
  "request_id": "req_translate_demo_001",
  "src": "en-Latn",
  "tgt": "fr-Latn",
  "content_mode": "plain",
  "preserve_placeholders": true,
  "do_not_translate": [
    "{name}"
  ],
  "metadata": {
    "order_id": "ord_123"
  },
  "input": {
    "items": [
      {
        "txt": "Open a support ticket for {name}."
      }
    ]
  }
}'

Response example

json

{
  "ok": true,
  "request_id": "req_translate_demo_001",
  "operation": "translate",
  "source_language_tag": "en-Latn",
  "source_language_name": "English",
  "detection_confidence": 0.99,
  "inference_latency": 142,
  "output": {
    "kind": "single",
    "single": {
      "translation": "Ouvrez un ticket d'assistance pour {name}.",
      "src": "en-Latn",
      "tgt": "fr-Latn"
    }
  },
  "usage": {
    "input_characters": 33,
    "output_characters": 42
  },
  "warnings": [],
  "quota": {
    "creditType": "premium",
    "membership": "premium",
    "monthlyCharacterQuota": 5000000,
    "monthlyCharacterUsage": 120033,
    "monthlyBillableCharacterUsage": 0,
    "perRequestCharacterLimit": null
  },
  "trial": {
    "isAuthenticated": true,
    "limit": null,
    "used": null,
    "remaining": null,
    "startedAt": null,
    "windowMs": null
  }
}

Request fields

Use one item in input.items for synchronous single translation. Root-level config keeps the public parser consistent with batch and async routes.

Field	Type	Required	Description
request_id	string	Yes	Caller-provided trace ID returned in responses and logs.
src	language tag \| ""	No	Root source language for the request. Use an empty string to auto-detect.
tgt	language tag	Yes	Root target language for the request.
content_mode	auto \| plain \| markdown \| html \| json \| xml \| yaml \| csv	No	Structured content is preprocessed before translation.
split_input	0 \| 1 \| 2	No	0 off, 1 balanced (default), 2 aggressive. Splitting can reduce latency with generally minimal impact on translation quality.
ps_context	string	No	Optional preceding-sentence context for short or ambiguous text.
metadata	object	No	Caller-defined JSON metadata echoed in logs and responses when present.
glossary_id	string	No	Stored glossary selection for this request.
glossary_version_id	string	No	Optional immutable stored glossary version for reproducible sync behavior.
glossary_group_ids	string[]	No	Selected enabled glossary groups. Combined with item selectors as a union.
glossary_item_ids	string[]	No	Selected enabled glossary items. Combined with group selectors as a union.
glossary.entries	array	No	Inline glossary mappings when you want request-scoped terminology instead of a stored glossary.
glossary.do_not_translate	string[]	No	Inline protected terms carried with the inline glossary.
do_not_translate	string[]	No	Top-level protected terms such as placeholders or brand names.
preserve_placeholders	boolean	No	Disabled by default. When true, the app detects placeholder-like patterns and adds them to protection rules without rewriting the source text.
strict_placeholder_preservation	boolean	No	Only applies when preserve_placeholders is true. Tightens enforcement when protected values are not preserved.
input.items	array(length = 1)	Yes	Native public single-translate requests use one item in the items array.
input.items[0].txt	string	Yes	Source text to translate.

Response fields

Field	Type	Required	Description
ok	boolean	Yes	True for successful translate responses.
request_id	string	Yes	Echoes the caller request ID or a generated request ID.
operation	"translate"	Yes	Response operation discriminator.
output.kind	"single"	Yes	Single-output discriminator.
output.single.translation	string	Yes	Canonical translated text field.
output.single.src	string	No	Canonical source language tag for the returned translation.
output.single.tgt	string	No	Canonical target language tag for the returned translation.
source_language_tag	string	No	Canonical supplied or detected source language tag.
source_language_name	string	No	Human-readable source language name.
detection_confidence	number \| null	No	Provider-backed detection confidence when available.
inference_latency	number \| null	No	Provider-reported inference latency in milliseconds when available.
usage.input_characters	number	Yes	Characters counted from the request input.
usage.output_characters	number	Yes	Characters in translated output.
warnings	array	Yes	Non-fatal warnings such as preservation or confidence notes.
quota	object \| null	No	Quota state returned for authenticated actors when available.
trial	object \| null	No	Runtime quota-window state returned in the current response envelope.

Errors

json

{
  "ok": false,
  "error": "Premium characters exhausted.",
  "status": 429,
  "request_id": "req_translate_demo_001",
  "quota": {
    "creditType": "premium",
    "membership": "premium",
    "monthlyCharacterQuota": 5000000,
    "monthlyCharacterUsage": 5000000,
    "monthlyBillableCharacterUsage": 0,
    "perRequestCharacterLimit": null
  }
}

Behavior

Structured content modes such as json and csv are preprocessed before translation. The provider sees extracted text segments, not your original structure.
Placeholder detection never rewrites the input text. When enabled, it only detects placeholder-like values and adds them to protected-term handling.

Next steps

See Synchronous batch for multiple items in one synchronous request.
See Glossaries for stored terminology selection.