API Reference - Abestar API

Explore every Abestar API endpoint and learn how to integrate them.

Public Metadata

Public

4 endpoints

Supported languages and models for translation and ASR; no auth.

Returns every language currently supported by the translation module, with code and display name. No auth header is required. The code follows BCP 47 (e.g. zh-TW, en-US) and can be passed directly as source_language / target_language on /api/v1/translations/text. Currently 27 languages are supported (including yue-Hant-HK, Cantonese (Hong Kong)).

Authorization

Public endpoint (no auth)

Request Examples

bash

curl https://abemono.abestar.com.tw/translations/languages

Responses

Full list of supported languages.

Name	Type	Description
languagesRequired	object[]	Array of language entries; each element `{ code: string, name: string }`. code is the BCP 47 tag, name is the English display name.

json

{
  "languages": [
    { "code": "en-US", "name": "English (US)" },
    { "code": "zh-TW", "name": "Traditional Chinese (Taiwan)" },
    { "code": "zh-CN", "name": "Simplified Chinese (China)" },
    { "code": "yue-Hant-HK", "name": "Cantonese (Hong Kong)" },
    { "code": "ja-JP", "name": "Japanese" }
  ]
}

Error Codes

internal_server_error

Notes

Translation and ASR modules maintain independent language lists. If you use both, call /translations/languages and /audio/languages separately to obtain each side's actual supported set.

Returns every translation model currently available, with code, display name, and description. No auth required. The code can be passed directly as the model field on /api/v1/translations/text.

Authorization

Public endpoint (no auth)

Request Examples

bash

curl https://abemono.abestar.com.tw/translations/models

Responses

Full list of available models.

Name	Type	Description
modelsRequired	object[]	Array of model entries; each element `{ code: string, name: string, description: string }`. code is the identifier passed to the translation endpoint.

json

{
  "models": [
    {
      "code": "PSSC-V2-251215",
      "name": "PSSC V2",
      "description": "第二代翻譯模型，改善翻譯品質"
    },
    {
      "code": "PSSC-V3-251215",
      "name": "PSSC V3",
      "description": "第三代翻譯模型，最佳翻譯品質"
    }
  ]
}

Error Codes

internal_server_error

Notes

This endpoint returns every model currently available for translation requests. Currently exposes 2 model codes (as listed above); the `description` field is written in Traditional Chinese.

Returns every language currently supported by the speech recognition module, with code and display name. No auth header is required. The code follows BCP 47 (e.g. zh-TW, en-US) and is used as the language field on /api/v1/audio/transcriptions. Currently 27 languages are supported (including yue-Hant-HK, Cantonese (Hong Kong)).

Authorization

Public endpoint (no auth)

Request Examples

bash

curl https://abemono.abestar.com.tw/audio/languages

Responses

Full list of supported languages.

Name	Type	Description
languagesRequired	object[]	Array of language entries; each element `{ code: string, name: string }`. code is the BCP 47 tag, name is the English display name.

json

{
  "languages": [
    { "code": "en-US", "name": "English (US)" },
    { "code": "zh-TW", "name": "Traditional Chinese (Taiwan)" },
    { "code": "zh-CN", "name": "Simplified Chinese (China)" },
    { "code": "yue-Hant-HK", "name": "Cantonese (Hong Kong)" },
    { "code": "ja-JP", "name": "Japanese" }
  ]
}

Error Codes

internal_server_error

Notes

The detected_language returned by /api/v1/audio/transcriptions is one of the codes in this list, or `"und"` (BCP-47 undetermined) when detection is inconclusive — "und" is not part of this list and cannot be used as the request's language value.

Returns every speech recognition model currently available, with code, display name, and description. No auth required. The code can be passed directly as the model field on /api/v1/audio/transcriptions.

Authorization

Public endpoint (no auth)

Request Examples

bash

curl https://abemono.abestar.com.tw/audio/models

Responses

Full list of available models.

Name	Type	Description
modelsRequired	object[]	Array of model entries; each element `{ code: string, name: string, description: string }`. code is the identifier passed to the ASR endpoint.

json

{
  "models": [
    {
      "code": "tranc-adv-v1",
      "name": "Advanced V1",
      "description": "進階辨識，適合品質需求較高的情境"
    },
    {
      "code": "tranc-pro-v1",
      "name": "Professional V1",
      "description": "高精度辨識，適合專業領域"
    }
  ]
}

Error Codes

internal_server_error

Notes

ASR and translation model lists are independent; do not assume the same code works on both endpoints. Currently exposes 2 model codes (as listed above); the `description` field is written in Traditional Chinese.