# Dropped Hints > A gift-list planner built on hints. Each person on the list is a gift tag that holds what they said, did and showed this year - one hint per line - plus a budget, what is ruled out (given before, already owned) and a status (thinking, got an idea, bought, wrapped). The tags, the clue sorting and the budget math are free in the browser; two metered lanes read one tag's hints into a gift idea, or plan a short conversation to find out what is still missing without spoiling the surprise. URL: https://dropped-hints.skillsafe.ai/ API tutorial: https://dropped-hints.skillsafe.ai/api.html ## Free, in the browser - The list as tags on a line of twine, coloured by status, each showing the budget, the paid amount or the picked idea's estimate, and the idea. - The ledger: people, total budget, paid so far, picked ideas (estimates, marked as such), what is left or over, how many are bought or wrapped, days to the giving date. - The back of the tag: every hint numbered (C1, C2 ...) and sorted onto an empathy map - Says, Thinks, Does, Feels - from its wording (a line starting says:, thinks:, does: or feels: is sorted as written). Direct asks ("I need a new one", "keeps eyeing", "broke") are marked ASK. - Tag checks: no name, no hints, fewer than three hints, no budget, empty quadrants, paid over budget. - Exports: the list as Markdown, CSV and JSON; printable tags; saved lists. ## Metered (gpt-terra), two tasks, one tag at a time - read: the hints -> the map (every hint placed once, with what it suggests), a verdict by ordered rules, ideas tied to hint ids with rough price estimates, things to avoid and things to find out. Verdicts, first match wins: wrap-it (a hint is a direct ask for one specific thing, not ruled out, within budget), fish-first (fewer than 3 hints, or no idea can be tied to two different hints), pick-one (two or three ideas, each tied to two hints or more, one picked). - fish: the things to find out -> a conversation plan: cover line, warm-up, one question per thing with what to listen for and a follow-up, a wrap-up, and what not to say. Casual when you ask the person yourself (no mention of gifts, the occasion or shopping), straight when you ask someone close to them. The model knows only the hints: it may not invent facts about the person, name brands or shops, or put prices in prose. The page checks every reply: hint ids exist, every hint is placed, the verdict follows the rules, prices sit inside the budget, nothing ruled out comes back, and a casual plan has no giveaway words. ## Sources Built on the SkillSafe skills @owl-listener/empathy-map - https://skillsafe.ai/skill/@owl-listener/empathy-map - and @owl-listener/interview-script - https://skillsafe.ai/skill/@owl-listener/interview-script Holiday shopping timing (43% of holiday shoppers plan to start before November): National Retail Federation, 2026 winter holiday outlook - https://nrf.com/research-insights/holiday-data-and-trends/winter-holidays