{ "ocr_languages": [ { "code": "afr", "name": "Afrikaans", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "amh", "name": "Amharic", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ara", "name": "Arabic", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "asm", "name": "Assamese", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "aze", "name": "Azerbaijani", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "aze_cyrl", "name": "Azerbaijani - Cyrillic", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "bel", "name": "Belarusian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ben", "name": "Bengali", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "bod", "name": "Tibetan", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "bos", "name": "Bosnian", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "bre", "name": "Breton", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "bul", "name": "Bulgarian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "cat", "name": "Catalan; Valencian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ceb", "name": "Cebuano", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ces", "name": "Czech", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "chi_sim", "name": "Chinese - Simplified", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "chi_tra", "name": "Chinese - Traditional", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "chr", "name": "Cherokee", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "cos", "name": "Corsican", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "cym", "name": "Welsh", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "dan", "name": "Danish", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "dan_frak", "name": "Danish - Fraktur (contrib)", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": false, "tessdata": false, "tessdata_best": false, "tessdata_fast": false }, { "code": "deu", "name": "German", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "deu_frak", "name": "German - Fraktur (contrib)", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": false, "tessdata": false, "tessdata_best": false, "tessdata_fast": false }, { "code": "deu_latf", "name": "German (Fraktur Latin)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "dzo", "name": "Dzongkha", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ell", "name": "Greek, Modern (1453-)", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "eng", "name": "English", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "enm", "name": "English, Middle (1100-1500)", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "epo", "name": "Esperanto", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "equ", "name": "Math / equation detection module", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "est", "name": "Estonian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "eus", "name": "Basque", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "fao", "name": "Faroese", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "fas", "name": "Persian", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "fil", "name": "Filipino (old - Tagalog)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "fin", "name": "Finnish", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "fra", "name": "French", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "frk", "name": "German - Fraktur (now deu_latf)", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "frm", "name": "French, Middle (ca.1400-1600)", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "fry", "name": "Western Frisian", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "gla", "name": "Scottish Gaelic", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "gle", "name": "Irish", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "glg", "name": "Galician", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "grc", "name": "Greek, Ancient (to 1453) (contrib)", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "guj", "name": "Gujarati", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hat", "name": "Haitian; Haitian Creole", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "heb", "name": "Hebrew", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hin", "name": "Hindi", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hrv", "name": "Croatian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hun", "name": "Hungarian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hye", "name": "Armenian", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "iku", "name": "Inuktitut", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ind", "name": "Indonesian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "isl", "name": "Icelandic", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ita", "name": "Italian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ita_old", "name": "Italian - Old", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "jav", "name": "Javanese", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "jpn", "name": "Japanese", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "kan", "name": "Kannada", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "kat", "name": "Georgian", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "kat_old", "name": "Georgian - Old", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "kaz", "name": "Kazakh", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "khm", "name": "Central Khmer", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "kir", "name": "Kirghiz; Kyrgyz", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "kmr", "name": "Kurmanji (Kurdish - Latin Script)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "kor", "name": "Korean", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "kor_vert", "name": "Korean (vertical)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "kur", "name": "Kurdish (Arabic Script)", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": false, "tessdata": false, "tessdata_best": false, "tessdata_fast": false }, { "code": "lao", "name": "Lao", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "lat", "name": "Latin", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "lav", "name": "Latvian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "lit", "name": "Lithuanian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ltz", "name": "Luxembourgish", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "mal", "name": "Malayalam", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "mar", "name": "Marathi", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "mkd", "name": "Macedonian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "mlt", "name": "Maltese", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "mon", "name": "Mongolian", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "mri", "name": "Maori", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "msa", "name": "Malay", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "mya", "name": "Burmese", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "nep", "name": "Nepali", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "nld", "name": "Dutch; Flemish", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "nor", "name": "Norwegian", "tesseract_3_02": true, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "oci", "name": "Occitan (post 1500)", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ori", "name": "Oriya", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "osd", "name": "Orientation and script detection module", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "pan", "name": "Panjabi; Punjabi", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "pol", "name": "Polish", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "por", "name": "Portuguese", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "pus", "name": "Pushto; Pashto", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "que", "name": "Quechua", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ron", "name": "Romanian; Moldavian; Moldovan", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "rus", "name": "Russian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "san", "name": "Sanskrit", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "sin", "name": "Sinhala; Sinhalese", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "slk", "name": "Slovak", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "slk_frak", "name": "Slovak - Fraktur (contrib)", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": false, "tessdata": false, "tessdata_best": false, "tessdata_fast": false }, { "code": "slv", "name": "Slovenian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "snd", "name": "Sindhi", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "spa", "name": "Spanish; Castilian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "spa_old", "name": "Spanish; Castilian - Old", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "sqi", "name": "Albanian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "srp", "name": "Serbian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "srp_latn", "name": "Serbian - Latin", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "sun", "name": "Sundanese", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "swa", "name": "Swahili", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "swe", "name": "Swedish", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "syr", "name": "Syriac", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "tam", "name": "Tamil", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "tat", "name": "Tatar", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "tel", "name": "Telugu", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "tgk", "name": "Tajik", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "tgl", "name": "Tagalog (new - Filipino)", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": false, "tessdata_best": false, "tessdata_fast": false }, { "code": "tha", "name": "Thai", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "tir", "name": "Tigrinya", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ton", "name": "Tonga", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "tur", "name": "Turkish", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "uig", "name": "Uighur; Uyghur", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ukr", "name": "Ukrainian", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "urd", "name": "Urdu", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "uzb", "name": "Uzbek", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "uzb_cyrl", "name": "Uzbek - Cyrillic", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "vie", "name": "Vietnamese", "tesseract_3_02": true, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "yid", "name": "Yiddish", "tesseract_3_02": false, "tesseract_3_04": true, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "yor", "name": "Yoruba", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": true, "tessdata": true, "tessdata_best": true, "tessdata_fast": true } ], "ocr_scripts": [ { "code": "arab", "name": "Arabic", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "armn", "name": "Armenian", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "beng", "name": "Bengali", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "cans", "name": "Canadian_Aboriginal", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "cher", "name": "Cherokee", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "cyrl", "name": "Cyrillic", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "deva", "name": "Devanagari", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "ethi", "name": "Ethiopic", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "frak", "name": "Fraktur", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "geor", "name": "Georgian", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "grek", "name": "Greek", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "gujr", "name": "Gujarati", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "guru", "name": "Gurmukhi", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hans", "name": "HanS (Han simplified)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hans-vert", "name": "HanS_vert (Han simplified vertical)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hant", "name": "HanT (Han traditional)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hant-vert", "name": "HanT_vert (Han traditional vertical)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hang", "name": "Hangul", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hang-vert", "name": "Hangul_vert (Hangul vertical)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "hebr", "name": "Hebrew", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "jpan", "name": "Japanese", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "jpan-vert", "name": "Japanese_vert (Japanese vertical)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "knda", "name": "Kannada", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "khmr", "name": "Khmer", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "laoo", "name": "Lao", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "latn", "name": "Latin", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "mlym", "name": "Malayalam", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "mymr", "name": "Myanmar", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "orya", "name": "Oriya(Odia)", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "sinh", "name": "Sinhala", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "syrc", "name": "Syriac", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "taml", "name": "Tamil", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "telu", "name": "Telugu", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "thaa", "name": "Thaana", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "thai", "name": "Thai", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "tibt", "name": "Tibetan", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true }, { "code": "viet", "name": "Vietnamese", "tesseract_3_02": false, "tesseract_3_04": false, "tesseract_4_nov_2016": false, "tessdata": true, "tessdata_best": true, "tessdata_fast": true } ], "transcription_languages": { "en": "english", "zh": "chinese", "de": "german", "es": "spanish", "ru": "russian", "ko": "korean", "fr": "french", "ja": "japanese", "pt": "portuguese", "tr": "turkish", "pl": "polish", "ca": "catalan", "nl": "dutch", "ar": "arabic", "sv": "swedish", "it": "italian", "id": "indonesian", "hi": "hindi", "fi": "finnish", "vi": "vietnamese", "he": "hebrew", "uk": "ukrainian", "el": "greek", "ms": "malay", "cs": "czech", "ro": "romanian", "da": "danish", "hu": "hungarian", "ta": "tamil", "no": "norwegian", "th": "thai", "ur": "urdu", "hr": "croatian", "bg": "bulgarian", "lt": "lithuanian", "la": "latin", "mi": "maori", "ml": "malayalam", "cy": "welsh", "sk": "slovak", "te": "telugu", "fa": "persian", "lv": "latvian", "bn": "bengali", "sr": "serbian", "az": "azerbaijani", "sl": "slovenian", "kn": "kannada", "et": "estonian", "mk": "macedonian", "br": "breton", "eu": "basque", "is": "icelandic", "hy": "armenian", "ne": "nepali", "mn": "mongolian", "bs": "bosnian", "kk": "kazakh", "sq": "albanian", "sw": "swahili", "gl": "galician", "mr": "marathi", "pa": "punjabi", "si": "sinhala", "km": "khmer", "sn": "shona", "yo": "yoruba", "so": "somali", "af": "afrikaans", "oc": "occitan", "ka": "georgian", "be": "belarusian", "tg": "tajik", "sd": "sindhi", "gu": "gujarati", "am": "amharic", "yi": "yiddish", "lo": "lao", "uz": "uzbek", "fo": "faroese", "ht": "haitian creole", "ps": "pashto", "tk": "turkmen", "nn": "nynorsk", "mt": "maltese", "sa": "sanskrit", "lb": "luxembourgish", "my": "myanmar", "bo": "tibetan", "tl": "tagalog", "mg": "malagasy", "as": "assamese", "tt": "tatar", "haw": "hawaiian", "ln": "lingala", "ha": "hausa", "ba": "bashkir", "jw": "javanese", "su": "sundanese", "yue": "cantonese" }, "sources": [ "https://tesseract-ocr.github.io/tessdoc/Data-Files-in-different-versions.html", "https://raw.githubusercontent.com/openai/whisper/main/whisper/tokenizer.py" ] }