I planned a walk in 60 seconds, then walked for two hours (open-weight AI, no GPU)

Описание и применение

Приложение позволяет пользователям за минуту получить план двухчасовой прогулки, включая маршрут, природный ориентир, список вещей и карту, используя открытый LLM без GPU и без платы.

Задача

People struggle to plan a quick outdoor walk that fits a tight time budget and provides useful details like a loop, nature cue, packing list, and weather.

Результат

The app delivers a one‑minute walk plan that users can act on immediately, with a 60:1 screen‑to‑outside time ratio. It runs for $0, uses no GPU, and remains functional despite provider changes.

Рабочий процесс

• 1. User enters location and time budget on the web form.
• 2. App queries Open‑Meteo for weather (cached).
• 3. App sends prompt to the LLM (openai/gpt‑oss‑20b) via Groq.
• 4. LLM returns walk plan: time window, loop, nature cue, packing list, map pin.
• 5. App displays plan and map pin.
• 6. User closes tab and walks.

Доказательства

One input gives you one walk: the best time window, a loop idea, a nature cue and a packing line, plus a live map pin.

Источник

On my own test walk I spent about one minute on the screen and one to two hours outside — roughly sixty to one.

Источник

The plan named a destination I would not have picked myself — the lake promenade — and the "no car needed" line meant I walked instead of driving somewhere and sitting in a car park.

Источник

Weather never rendered on the deployed app. Render's free tier shares one outbound IP across many instances, and Open‑Meteo rate‑limits that IP.

Источник