Yes. Add your rate matrix as a pricing page or a structured table. SleekAI reads it as context and quotes the right per-word rate for each pair. Spanish-English, Mandarin-English, Russian-English, Arabic-English, French-English, and any other pair you handle get their own rate. The bot also handles minimum charges, page-based versus word-based pricing, and domain factors (legal, medical) as separate inputs that combine into the final quote on every conversation.

Each format gets its own page: USCIS-format certified, court-format sworn translation, academic credential evaluation, and corporate certified. The bot reads each and applies the right one based on the prospect's stated use case. USCIS specifically requires a Certificate of Translation Accuracy and does not require notarisation. State courts often require a sworn translator's affidavit and may require notarisation. Academic credential evaluations need WES or ECE-compatible formatting. Corporate contracts usually want translator credentials on the cover. The bot distinguishes them so the deliverable matches the filing.

Yes, when those are surcharge lines on your pricing page. Rush turnaround triggers a confirmation that a specialist in the relevant language pair is available before the bot commits to a deadline. Notarisation only applies when the prospect's filing requires it (USCIS does not; state courts often do; international filings under the Hague Convention need apostille instead). Apostille is a separate filing with the Secretary of State and gets quoted as a separate line item per your published apostille service page.

Yes. Translator profile pages live as a custom post type with fields for language pair, specialty (legal, medical, technical, marketing, patent, literary), and certifications. The bot reads the directory and recommends the best-fit specialist based on the project type. Legal contracts go to a legal specialist who also handles the certified-translation policy. Medical records go to a medical specialist with relevant terminology experience. Patent translations route to a translator with relevant patent-office experience. The match happens before the project manager picks it up.

Yes, when you have a localisation service page. Website localisation has different inputs from document translation: word count comes from a crawl rather than a file, file format is usually XLIFF or CSV, the project may need translation-memory leveraging, and SEO factors (keyword adaptation, slug translation) often apply. The bot captures URL, target languages, scope, CMS, and translation-memory needs, then quotes from your localisation rate table. For very large projects (50,000+ words), the bot routes to a project manager for a custom estimate rather than committing to a number inline.

The system prompt instructs Verba to capture project metadata (language pair, word count, deadline, certification format) rather than document content. Actual files move through your secure portal after the project is logged, never pasted into chat. This matters for legal contracts, medical records, and corporate documents where the client expects NDA-grade handling. The bot can be told explicitly to refuse if a prospect tries to paste a confidential document into the chat window, and the guideline filter enforces that refusal even when the system prompt is bypassed.

SleekAI does not connect directly to memoQ, Trados, Smartcat, or Phrase. It logs the project intake in WordPress, and the project manager pushes the summary into the TMS through a webhook or a manual handoff. Most agencies use a small Zapier or Make flow to forward intake summaries from the SleekAI webhook into their TMS project queue, which avoids heavy custom-integration work. The translation memory itself stays in the TMS, where it should; the bot's job ends at intake.

Yes. GPT-4o, Claude, and Gemini all handle most major languages natively in conversation, including domain-specific vocabulary in Spanish, Mandarin, Russian, Portuguese, Arabic, French, and German. Many agencies run a single multilingual bot rather than one bot per language. Prospects often start in their source language and switch to English mid-conversation, and the bot handles the switch seamlessly. Native-speaker translators still do the actual translation work, of course; the bot just handles intake and quoting.