Language Identification API

Identify the language of text using Gherbal, Sawalni's fastText-based language identification model. Supports 200+ languages with ISO 639-3 + script codes.

Endpoint

text

POST https://api.sawalni.com/v1/language/identify

Request Body

Parameter	Type	Required	Description
`model`	string	Yes	`gherbal-v4`
`input`	string or array	Yes	Text(s) to identify
`top`	integer	No	Number of predictions per text. Default: 1. Use -1 for all.

Example — curl

bash

curl https://api.sawalni.com/v1/language/identify \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gherbal-v4",
    "input": "مرحبا بالعالم",
    "top": 3
  }'

Example — Python

python

import requests

response = requests.post(
    "https://api.sawalni.com/v1/language/identify",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={
        "model": "gherbal-v4",
        "input": ["مرحبا بالعالم", "Hello world", "Bonjour le monde"],
        "top": 3
    }
)

for item in response.json()["data"]:
    top = item["predictions"][0]
    print(f"{item['input']} → {top['language']} ({top['confidence']:.2%})")

Example — JavaScript

javascript

const response = await fetch('https://api.sawalni.com/v1/language/identify', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer YOUR_API_KEY',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    model: 'gherbal-v4',
    input: ['مرحبا', 'Hello', 'Bonjour'],
    top: 3
  })
});

const result = await response.json();
result.data.forEach(item => {
  console.log(item.input, '→', item.predictions[0].language);
});

Response

json

{
  "data": [
    {
      "input": "مرحبا بالعالم",
      "predictions": [
        { "language": "arb_Arab", "confidence": 0.987 },
        { "language": "arz_Arab", "confidence": 0.008 },
        { "language": "acm_Arab", "confidence": 0.003 }
      ]
    }
  ],
  "model": "gherbal-v4",
  "usage": {
    "input_bytes": 25,
    "input_count": 1,
    "processing_ms": 2.1
  }
}

Language Code Format

Labels follow the ISO 639-3 + Unicode script format: {iso639-3}_{script}

Examples: arb_Arab (Modern Standard Arabic), eng_Latn (English), fra_Latn (French), deu_Latn (German), zho_Hans (Chinese Simplified)

Batch Identification

Pass an array of strings to identify multiple texts in one request (max 1024):

python

response = requests.post(
    "https://api.sawalni.com/v1/language/identify",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={
        "model": "gherbal-v4",
        "input": ["First text", "Second text", "Third text"],
        "top": 1
    }
)

The `top` Parameter

Value	Behavior
`1`	Returns only the most likely language (default)
`3`	Returns top 3 predictions
`10`	Returns top 10 predictions
`-1`	Returns all supported languages ranked by confidence

Models

Model	Description
`gherbal-v4`	Language identification for 200+ languages

Pricing

Language identification is metered by input bytes (UTF-8 encoded size).

Metric	Rate
Input bytes	0.05 credits per 1M bytes

Use Cases

Content routing — Route user messages to language-specific pipelines
Data quality — Filter datasets by language
Multilingual UX — Auto-detect user's language for localization
Moderation — Identify language before applying language-specific filters

Parameter

Type

Required

Description

model

string

Yes

gherbal-v4

input

string or array

Yes

Text(s) to identify

top

integer

Number of predictions per text. Default: 1. Use -1 for all.

curl https://api.sawalni.com/v1/language/identify \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "gherbal-v4", "input": "مرحبا بالعالم", "top": 3 }'

Example — Python

python

import requests

response = requests.post(
    "https://api.sawalni.com/v1/language/identify",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={
        "model": "gherbal-v4",
        "input": ["مرحبا بالعالم", "Hello world", "Bonjour le monde"],
        "top": 3
    }
)

for item in response.json()["data"]:
    top = item["predictions"][0]
    print(f"{item['input']} → {top['language']} ({top['confidence']:.2%})")

Example — JavaScript

javascript

const response = await fetch('https://api.sawalni.com/v1/language/identify', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer YOUR_API_KEY',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    model: 'gherbal-v4',
    input: ['مرحبا', 'Hello', 'Bonjour'],
    top: 3
  })
});

const result = await response.json();
result.data.forEach(item => {
  console.log(item.input, '→', item.predictions[0].language);
});

Response

json

{
  "data": [
    {
      "input": "مرحبا بالعالم",
      "predictions": [
        { "language": "arb_Arab", "confidence": 0.987 },
        { "language": "arz_Arab", "confidence": 0.008 },
        { "language": "acm_Arab", "confidence": 0.003 }
      ]
    }
  ],
  "model": "gherbal-v4",
  "usage": {
    "input_bytes": 25,
    "input_count": 1,
    "processing_ms": 2.1
  }
}

Batch Identification

Pass an array of strings to identify multiple texts in one request (max 1024):

python

response = requests.post(
    "https://api.sawalni.com/v1/language/identify",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={
        "model": "gherbal-v4",
        "input": ["First text", "Second text", "Third text"],
        "top": 1
    }
)

Value

Behavior

1

Returns only the most likely language (default)

3

Returns top 3 predictions

10

Returns top 10 predictions

-1

Returns all supported languages ranked by confidence

Model

Description

gherbal-v4

Language identification for 200+ languages

Metric

Rate

Input bytes

0.05 credits per 1M bytes

Language Identification API

Endpoint

Request Body

Example — curl

Example — Python

Example — JavaScript

Response

Language Code Format

Batch Identification

The top Parameter

Models

Pricing

Use Cases

Language Identification API

Endpoint

Request Body

Example — curl

Example — Python

Example — JavaScript

Response

Language Code Format

Batch Identification

The top Parameter

Models

Pricing

Use Cases

The `top` Parameter

The `top` Parameter