6 languages c distinct coverage profiles. Hebrew RTL, Russian Cyrillic, English production-GA across all 9 verticals; German / French beta; Arabic Q4 2026 roadmap. Differentiator vs auto-translate-only platforms — SLAtech ships first-class native rendering, tokenization и NER per language.
Coverage: First-class. All UI strings, FAQ content, system prompts, error messages, email templates.
Default fallback language. Used as anchor for the multilingual sitemap (hreflang x-default).
Coverage: First-class. RTL UI rendering, Hebrew-tuned tokenization, font choice (Heebo / Assistant), date / number / currency formatting per Israeli locale.
Polished RTL — admin platform, widget UI и client-facing emails all render RTL correctly. Differentiator: Intercom / Drift / Chatbase / Tidio Lyro ship LTR-only UIs с Hebrew text inserted.
Coverage: First-class. Cyrillic-aware tokenization, Cyrillic-aware NER, full Russian UI strings across admin и widget.
Russian-speaking Israeli demographics (~1.5M speakers) и EU Russian-speaking diaspora are а significant SLAtech audience. Coverage exceeds platforms targeting US-only markets.
Coverage: RTL UI rendering, Arabic-tuned tokenization, locale formatting per UAE / Saudi standards. Roadmap delivery target Q4 2026.
GCC market entry. Will ship с both Modern Standard Arabic (MSA) и Levantine dialect support для customer-facing conversations.
Coverage: Beta — translatable via the multilingual pipeline; admin UI partially translated; widget client-facing translation production-ready.
DACH market entry. Awaiting full German legal-language review (GDPR-, MiFID II- adjacent compliance terminology).
Coverage: Beta — same status as German.
FR market entry. CNIL-conformant data-handling language pending review.
Each incoming visitor message routes through а 3-stage detection: (1) explicit lang query param (если widget initialised с а lang), (2) accept-language HTTP header (browser default), (3) statistical detection from message content (cld3 + custom Hebrew / Cyrillic / Arabic detectors). The detected language pins the LLM call's system prompt и retrieval-pipeline filter.
Tenant knowledge content is indexed per language. Query in Hebrew → retrieves Hebrew chunks first (с а fallback к English if scores below threshold). Prevents the "answer in English when the question was in Hebrew" failure mode that plagues auto-translate-based platforms.
Bidirectional Algorithm-correct rendering: numbers in Hebrew text use weak directionality, currency symbols use strong directionality, mixed Hebrew/English use explicit segmenting. Widget UI tested на 6 desktop browsers + 4 mobile browsers. Email templates use direction: rtl; explicitly. Admin platform respects user-language-preference saved per-account.
Cyrillic-aware tokenization preserves Russian morphology better than naive whitespace-tokenization. Custom NER tuned to recognise Russian-named entities (companies, products, locations). Date / number / currency formatting follows Russian conventions (DD.MM.YYYY, space-separated thousands).
А single tenant can have visitors arriving в different languages. The bot detects the visitor's language и serves а matched-language conversation. Admin Inbox surfaces а language tag per conversation для analytics и agent assignment.
When tenant content is missing в the visitor's detected language, SLAtech can optionally serve auto-translated content with а disclosure ("This response was auto-translated from English"). Tenant admin controls whether auto-translate is enabled. Off by default для compliance-sensitive verticals (Med, Legal).
SLAtech is the only first-class Hebrew-RTL и Russian production-grade chatbot platform. Try the trial.