Loading page...Загружаем страницу...正在加载页面...
AtlasRepo

TouchGrass AI: I built an outdoor quest app with Ollama and local-first storage

Description and use

TouchGrass AI is an outdoor quest companion that uses an open AI model (Ollama/Gemma) to generate personalized quests and stores all data, including photos, locally in the browser, enabling offline use and privacy protection.

Problem

Users want an outdoor quest companion that generates personalized quests without relying on continuous internet or uploading photos, preserving privacy and offline usability.

Outcome

TouchGrass AI delivers a privacy‑preserving, offline‑capable outdoor quest app that generates quests with an open model, stores data locally, and handles photo evidence without uploading.

Workflow

  • 1. User selects activity type, duration, and difficulty in the UI.
  • 2. UI sends selection to same‑origin server route.
  • 3. Server calls Ollama (local or API) to generate structured JSON.
  • 4. Server validates JSON against quest rules; if invalid or generation fails, returns a ready‑made quest.
  • 5. Client receives quest, stores it and any photo evidence in IndexedDB.
  • 6. User can view, edit, or delete quests offline; photos are previewable and removable but never cached or uploaded.
  • 7. App shell loads offline after first visit, allowing saved quests to be used without internet.

Evidence

The app keeps quests, progress, and photo evidence in the browser on that device. Photos are stored as blobs in IndexedDB, can be previewed or removed, and never go into the service worker cache.

Source

The UI sends the selected activity, duration, and difficulty to a same-origin server route. That route asks Ollama for structured JSON, then validates the response against the quest rules before showing it.

Source

For local development I run Ollama on my computer. For the hosted demo I configured Ollama's API with a Gemma model. The API key stays in the server environment; it is not included in browser code.

Source

The mobile test confirmed the camera/photo-picker flow works, and the evidence remains attached to the saved quest on that device.

Source

  • Record date: 2026-10-06T22:41:02.185Z