italki language learning

The target language the user wants to learn or practice (single value, lowercase English). Example: "I want to practice Spanish in group learning" → "spanish"

**REQUIRED.** The locale/language code of the language the user is currently using to communicate with you (GPT). You must detect this from the user's input and return it as a lowercase locale code (e.g. "en", "zh-cn", "zh-tw", "es", "fr", "de", "ja", "ko").

Optional. Categories of interest. If you are unable to determine the user's interest category, you don't need to give a value. If you are clear about the interest, must use exact values from this list: "Daily conversation", "Language essentials", "Travel and global culture", "Business and career", "Debating and discussion", "Hobbies and Interests", "Exam preparation".

Optional. The user intent of the current conversation session, **must be in English**, within 60 characters. **IMPORTANT:** Extract the specific, accurate intent based on the conversation context - only provide this field when you have enough information (90%+ confidence) to accurately reflect what the user is trying to achieve.

Optional. The user's proficiency level in the target language. Values: "beginner", "intermediate","advanced". Example: "I'm a beginner in Spanish" → "beginner"

Optional. The purpose of learning the target language, **must be in English**, within 60 characters. **IMPORTANT:** Extract the specific, accurate purpose based on sufficient context - only provide this field when you have enough information (90%+ confidence) to accurately understand the user's learning goals.

The target language the user wants to test (e.g., "English", "Spanish", "Japanese"). Example: "learning Spanish" → "spanish"

**REQUIRED.** The locale/language code of the language the user is currently using to communicate with you (GPT). You must detect this from the user's input and return it as a lowercase locale code (e.g. "en", "en-us", "zh-cn", "zh-tw", "es", "fr", "de", "ja", "ko", "ar", "th", "vi", "ru", "pt", "it", "tr", "pl", "nl"). This follows the same format as `window.openai.locale` (lowercase). **Important:** (1) For Chinese: use "zh-cn" for Simplified Chinese, "zh-tw" for Traditional Chinese (Taiwan/Hong Kong). (2) For English: use "en" or "en-us" (both will be normalized to "en"). (3) For other languages: use the ISO 639-1 language code (e.g. "es" for Spanish, "fr" for French, "ja" for Japanese, "ko" for Korean). This is NOT the target test language, but the language the user is speaking/writing in. **You must always provide this parameter based on the actual language of the user's message.** Example: If the user writes in Simplified Chinese, return "zh-cn"; if the user writes in English, return "en" or "en-us"; if the user writes in Spanish, return "es".

Optional. The user intent of the current conversation session, **must be in English**, within 60 characters. **IMPORTANT:** Extract the specific, accurate intent based on the conversation context - only provide this field when you have enough information (90%+ confidence) to accurately reflect what the user is trying to achieve. **Example:** `"I want to test my English level for job applications"` → `"Test English proficiency for job applications"` (NOT just "Test English")

Optional. The purpose of learning/testing the target language, **must be in English**, within 60 characters. **IMPORTANT:** Extract the specific, accurate purpose based on sufficient context - only provide this field when you have enough information (90%+ confidence) to accurately understand the user's goals. **Example:** `"I need to pass IELTS exam with band 7 for university application"` → `"Pass IELTS band 7 for university application"` (NOT just "Pass IELTS")

Target language (lowercase English, e.g. "spanish", "english", "chinese"). Example: "learning Spanish" → "spanish"

**REQUIRED.** The locale/language code of the language the user is currently using to communicate with you (GPT). You must detect this from the user's input and return it as a lowercase locale code (e.g. "en", "en-us", "zh-cn", "zh-tw", "es", "fr", "de", "ja", "ko", "ar", "th", "vi", "ru", "pt", "it", "tr", "pl", "nl"). This follows the same format as `window.openai.locale` (lowercase). **Important:** (1) For Chinese: use "zh-cn" for Simplified Chinese, "zh-tw" for Traditional Chinese (Taiwan/Hong Kong). (2) For English: use "en" or "en-us" (both will be normalized to "en"). (3) For other languages: use the ISO 639-1 language code (e.g. "es" for Spanish, "fr" for French, "ja" for Japanese, "ko" for Korean). This is NOT the target learning language, but the language the user is speaking/writing in. **You must always provide this parameter based on the actual language of the user's message.** Example: If the user writes in Simplified Chinese, return "zh-cn"; if the user writes in English, return "en" or "en-us"; if the user writes in Spanish, return "es".

Optional. Languages teachers also speak (array, lowercase English). Map 'mandarin' to 'chinese' if needed. Example: ['french', 'chinese']

Optional. The user intent of the current conversation session, **must be in English**, within 60 characters. **IMPORTANT:** Extract the specific, accurate intent based on the conversation context - only provide this field when you have enough information (90%+ confidence) to accurately reflect what the user is trying to achieve. **Example:** `"I'm beginner in Italian, how can I learn Italian with natives living in Italy?"` → `"Find Italian tutors who are natives living in Italy"` (NOT just "Learn Italian")

Optional. If the user's learning purpose is to prepare for an exam, include **only** the following predefined values of the exam name (array of exam codes). Chinese exams: `"HSK"`. English exams: `"TOEFL"`, `"TOEIC"`, `"IELTS"`, `"FCE"`, `"BEC"`, `"PET"`, `"CAE"`, `"CPE"`, `"KET"`, `"ILEC"`, `"OET"`. Japanese exams: `"EJU"`, `"JLPT"`. Spanish exams: `"CELU"`, `"DELE"`. Korean exams: `"KLPT"`, `"TOPIK"`. Italian exams: `"PLIDA"`, `"CILS"`, `"CELI"`. German exams: `"TestDaF"`, `"DSH"`. French exams: `"DELF"`, `"TELC"`, `"TEF"`, `"TCF"`. Portuguese exams: `"CELPE-Bras"`. Russian exams: `"TORFL"`. Arabic exams: `"ALPT"`. **Example:** `["IELTS", "TOEFL"]`

Optional. Teachers' origin countries (array of ISO country codes, e.g. ['GB', 'US', 'JP']). Extract country for accent learning. Example: 'British accent' → ['GB']

Optional. Filter teachers by whether they are native speakers. 1 indicates native speaker, 0 or undefined indicates any teacher.

Optional. The purpose of learning the target language, **must be in English**, within 60 characters. **IMPORTANT:** Extract the specific, accurate purpose based on sufficient context - only provide this field when you have enough information (90%+ confidence) to accurately understand the user's learning goals. **Example:** `"How can I get score 7 in the IELTS exam in 2 months? My last score was 6.5, my weakness is speaking and writing."` → `"Improve IELTS speaking and writing from 6.5 to band 7"` (NOT just "Prepare for IELTS")

Optional. Maximum price filter in **cents (USD)**. Extract **only the numeric value** from the price mentioned by the user, **without any currency symbol or unit**. **Important:** (1) The value must be in cents (USD) as an integer. (2) If the user inputs prices in other currencies (e.g., CNY, EUR, GBP), you must convert them to USD cents using the current exchange rate and round to the nearest integer. (3) If the user already mentions USD/dollar prices, convert dollars to cents (multiply by 100). (4) Recognize various price expressions: "不超过", "不超过", "less than", "below", "under", "up to", "maximum", "at most", "no more than", etc. (5) **If the user mentions a price range** (e.g., "5 to 10", "between 5 and 10", "5-10"), **extract the maximum value** from the range for max_price. (6) If only one value is given, fill the corresponding field and leave the other empty. **Example:** `"价格不超过250元"` → convert CNY to USD → convert USD to cents → `"max_price": "..."`, `"价格5到10元"` → `"max_price": "..." (10 converted)`, `"价格大于5元小于250元"` → `"min_price": "...", "max_price": "..."`, `"under $20"` → `"max_price": "2000"`

Optional. Minimum price filter in **cents (USD)**. Extract **only the numeric value** from the price mentioned by the user, **without any currency symbol or unit**. **Important:** (1) The value must be in cents (USD) as an integer. (2) If the user inputs prices in other currencies (e.g., CNY, EUR, GBP), you must convert them to USD cents using the current exchange rate and round to the nearest integer. (3) If the user already mentions USD/dollar prices, convert dollars to cents (multiply by 100). (4) Recognize various price expressions: "大于", "more than", "above", "at least", "from", "starting at", etc. (5) **If the user mentions a price range** (e.g., "5 to 10", "between 5 and 10", "5-10"), **extract the minimum value** from the range for min_price. (6) If only one value is given, fill the corresponding field and leave the other empty. **Example:** `"价格大于5元"` → convert CNY to USD → convert USD to cents → `"min_price": "..."`, `"价格5到10元"` → `"min_price": "..." (5 converted)`, `"价格大于5元小于250元"` → `"min_price": "...", "max_price": "..."`, `"$10 minimum"` → `"min_price": "1000"`