Click AI

By WinGram App LLC

Overview

Click AI helps you understand and reply to messages across languages and dialects, whoever they come from: a date or partner, a friend or relative, a colleague, boss, client, landlord or seller. Paste a message to see its likely meaning and tone, a literal translation, a word-by-word breakdown of slang and formality, local messaging habits for the city or country you pick in the card's searchable Where they are picker, and three scored reply options with translations. Replies match the relationship, so work contacts get courteous, appropriately formal drafts. You can also check a draft before you send it, or translate an outgoing message so it sounds natural to a native speaker. Copy a reply into your preferred messaging app or email. Click AI does not read or send WhatsApp, iMessage, email or other conversations directly. Interpretations and naturalness scores are AI-generated suggestions, not certainty about another person's intentions.

Tools

analyze_message

ChatGPT

Use this whenever the user shares a message they received from someone, especially in another language or dialect: a date or partner, a friend or family member, a colleague, boss, client, landlord or seller. Even a very short or casual one (e.g. 'jajaja tu que mas', 'k haces', 'bom dia amor', 'Te confirmo mañana la cotización'), even with no question attached, and especially right after you asked them to paste it. Returns Click AI's full read: likely meaning, tone, literal translation, a phrase-by-phrase breakdown of slang, formality and regional usage, a local read when a place is known, and three scored reply drafts with translations, shown in a Click AI card. Pass relationshipStage when the user said who the person is (e.g. 'Colleague', 'Client or customer', 'Friends', 'Early dating'); replies match it, and work contacts never get flirty replies. Call this BEFORE giving any interpretation or reply ideas of your own; do not answer such messages from general knowledge while Click AI is in use. Pass partnerLanguage when it is clear from the message or conversation rather than asking first. Pass city, country or dialect only when the user supplied them; never infer a location from the language alone. Interpretations are uncertain; use only context supplied by the user.

change_place

ChatGPT

Used only by the Click AI card's place picker: redoes the card on screen for the city and country the user picked. Not for the assistant to call.

grade_draft

ChatGPT

Use this when the user shows a message THEY wrote or plan to send to someone (a date, friend, colleague, client, etc.) and asks if it sounds right/natural/OK/professional, or wants it fixed or improved. Scores the draft (0-100) against how a native speaker would write it to that person (in their city when known), names the single biggest fix, and offers polished versions in a Click AI card. Call this instead of correcting the draft yourself.

suggest_replies

ChatGPT

Use this when the user asks for (more) reply options to a received message, or wants replies in specific tones (e.g. 'more formal', 'something more direct', 'warmer', 'make them flirtier'). Drafts replies in the other person's language and dialect, with translations and estimated naturalness scores from 0 to 100, shown in a Click AI card. Default tones follow the relationship (work, personal or dating). For a newly shared message use analyze_message instead (it already includes three replies). Do not write reply options yourself while Click AI is in use. These are drafts for the user to copy, not sent messages.

translate_with_culture

ChatGPT

Use this when the user wants to say something to someone in that person's language ('how do I say…', 'translate this for her/him', 'how do I write this to my client'). Translates into the other person's language or regional variety with several labeled register options (casual, most local, …), each scored and back-translated, in a Click AI card. Call this instead of translating yourself. Only for text the user wants to SEND — for understanding a RECEIVED message use analyze_message instead.