Text & Number Toolkit
Pricing
from $1.00 / 1,000 tool calls
Text & Number Toolkit
Amounts and numbers in words in 56 languages with correct Arabic cheque grammar; CLDR number formatting and parsing (lakh grouping, Arabic-Indic digits); digit conversion, Arabic normalisation, transliteration, script detection; date parsing in 200 locales + Hijri. MCP server + REST. No logging.
Pricing
from $1.00 / 1,000 tool calls
Rating
0.0
(0)
Developer
Mohamad Najia
Maintained by CommunityActor stats
0
Bookmarked
1
Total users
0
Monthly active users
9 hours ago
Last modified
Categories
Share
Text & Number Toolkit — amounts in words, locale numbers, Arabic text, human dates — for AI agents
MCP server + REST API. Ten deterministic tools for the multilingual chores agents fumble: writing AED 1,250.50 on a cheque in Arabic with the right grammar (فقط ألف ومئتان وخمسون درهماً وخمسون فلساً لا غير), formatting ₹12,34,567.89 the Indian way, reading 1.234,56 as a German reader would, turning ١٥ رمضان ١٤٤٧ or 3 mars 2026 into an ISO date, cleaning Arabic text for search, and romanising any script to ASCII.
No AI inside — rules and Unicode/CLDR data — so results are exact, repeatable and cheap. Every answer states its confidence, what was assumed, and the standard behind it. Nothing is logged.
Tools (10)
| tool | what | price |
|---|---|---|
amount_to_words | money in words, 56 languages; Arabic with number–noun agreement and the right minor unit for 28 currencies; cheque style | $0.001 |
number_to_words | cardinal / ordinal / year in 56 languages | free* |
format_number | CLDR formatting: decimal, currency, percent; lakh grouping, Arabic-Indic digits, symbol placement | $0.001 |
parse_number | read 1.234,56, ١٬٢٣٤٫٥٦, 12 345,6 into an exact decimal; flags heuristic reads | $0.001 |
convert_digits | Western ↔ Arabic-Indic ↔ Persian digits inside any text | free* |
normalize_arabic | strip tashkeel/tatweel, unify alef, optional ة→ه ى→ي, digits, punctuation; reports what changed | $0.001 |
transliterate | any script → ASCII (slugs, matching); Arabic names flagged low-confidence | $0.001 |
detect_script | scripts, direction, mixed-script flag, digit systems, Persian/Urdu hints | free |
parse_date | human dates in 200 locales + relative phrases + Hijri (Umm al-Qura) → ISO, with precision and language | $0.002 |
list_languages | supported languages and curated currencies | free |
* free on the self-hosted endpoint without a key (rate-limited); metered on Apify.
Why the Arabic amount grammar matters
Arabic counted nouns change form with the number: درهم واحد (1), درهمان (2), ثلاثة دراهم (3–10), أحد عشر درهماً (11–99), مائة درهم (100), مئتا دينار (200). Generic libraries get this wrong, and a cheque or contract with the wrong form looks amateurish at best. This toolkit encodes the rule and the correct minor units (فلس، هللة، قرش، بيسة…) for the GCC, Levant, North Africa and major world currencies; anything else falls back to the CLDR name and says so.
Example
// parse_date {"text": "١٥ رمضان ١٤٤٧"}{"result": {"input": "١٥ رمضان ١٤٤٧", "calendar": "hijri", "hijri": "1447-09-15", "date": "2026-03-04", "weekday": "Wed", "language": "ar"},"data_confidence": "medium","estimated_items": ["Hijri→Gregorian by the Umm al-Qura calendar (±1 day vs. local moon sighting)"],"sources": ["https://github.com/dralshehri/hijridate"], "data_version": "2026.10.31-1", "warnings": [], "disclaimer": null}
Connect
- Apify (metered): Standby URL +
/mcp, headerAuthorization: Bearer <your Apify token>. - Self-hosted:
claude mcp add textkit --transport http https://textkit.noahsarch.tech/mcp --header "X-API-Key: <key>"— one Noah's Arch key works for all our agents. - REST:
POST https://textkit.noahsarch.tech/v1/tools/<tool>; OpenAPI at/openapi.json, docs at/docs.
Built on
num2words (56 languages), Babel (CLDR), dateparser (200 locales), hijridate (Umm al-Qura), anyascii, plus our own Arabic grammar and normalisation layer. All pure Python; no network at request time.
بالعربية
عدّة النصوص والأرقام — عشر أدوات حتمية لوكلاء الذكاء الاصطناعي: كتابة المبالغ بالحروف في 56 لغة مع قواعد العدد والمعدود الصحيحة بالعربية (درهم واحد، درهمان، ثلاثة دراهم، أحد عشر درهماً) والوحدات الصغرى الصحيحة (فلس، هللة، قرش…) للشيكات والعقود؛ تنسيق الأرقام وقراءتها حسب كل منطقة؛ تحويل الأرقام الهندية/اللاتينية؛ تطبيع النص العربي للبحث؛ النقحرة إلى ASCII؛ وقراءة التواريخ المكتوبة بلغة بشرية في 200 لغة بما فيها التواريخ الهجرية. لا ذكاء اصطناعي داخلها ولا تسجيل للاستعلامات.