Language translation

The translation service converts text between language pairs using neural machine translation models, running fully on the Qualcomm NPU. Batch translation is supported — multiple strings can be translated in a single request.

Prerequisites

Ensure the Audio Analytics container is deployed and running. See Running the Audio Analytics container for setup instructions.

Available models

Query the models endpoint to get the list of available translation models:

curl http://localhost:8085/audio-analytics/v1/api/translations/models

The currently available models are:

Model	Translation
`opus-en-es`	English → Spanish
`opus-es-en`	Spanish → English
`opus-en-zh`	English → Chinese
`opus-zh-en`	Chinese → English

Use the name value from the response in your translation requests.

Endpoints

Method	Endpoint	Description
`GET`	`/translations/models`	List available translation models
`POST`	`/translations/translate`	Translate text

Translate text

curl -X POST http://localhost:8085/audio-analytics/v1/api/translations/translate \
  -H "Content-Type: application/json" \
  -d '{
    "text": ["Hello world", "How are you?"],
    "model": "opus-en-es",
    "source_language": "en",
    "target_language": "es"
  }'

Parameters:

text (required): Array of strings to translate
model (required): Translation model name
source_language (required): Source language code matching the selected model (e.g. en for opus-en-es)
target_language (required): Target language code

Response:

{
  "translations": [
    {
      "translated_text": "Hola mundo",
      "source_language": "en",
      "target_language": "es"
    },
    {
      "translated_text": "¿Cómo estás?",
      "source_language": "en",
      "target_language": "es"
    }
  ]
}

Batch translation

Translate multiple strings in a single request by passing an array of texts:

curl -X POST http://localhost:8085/audio-analytics/v1/api/translations/translate \
  -H "Content-Type: application/json" \
  -d '{
    "text": ["Hello world", "How are you?", "Good morning"],
    "model": "opus-en-es",
    "source_language": "en",
    "target_language": "es"
  }'

Response:

{
  "translations": [
    {"translated_text": "Hola mundo", "source_language": "en", "target_language": "es"},
    {"translated_text": "¿Cómo estás?", "source_language": "en", "target_language": "es"},
    {"translated_text": "Buenos días.", "source_language": "en", "target_language": "es"}
  ]
}

Results are returned in the same order as the input array.

Python example

import requests

def translate(texts, model="opus-en-es", source_language="en", target_language="es"):
    url = "http://localhost:8085/audio-analytics/v1/api/translations/translate"
    response = requests.post(url, json={
        "text": texts,
        "model": model,
        "source_language": source_language,
        "target_language": target_language
    }, timeout=30)
    response.raise_for_status()
    return [t["translated_text"] for t in response.json()["translations"]]

results = translate(["Hello world", "How are you?"])
print(results)

IQ-9075 EVK

Prerequisites

Available models

Endpoints

Translate text

Batch translation

Python example

Next steps

Speech recognition

Text-to-speech

​Prerequisites

​Available models

​Endpoints

​Translate text

​Batch translation

​Python example

​Next steps

Speech recognition

Text-to-speech

Prerequisites

Available models

Endpoints

Translate text

Batch translation

Python example

Next steps