An offline trail planner built with Llama 3.2, Ollama, and Streamlit for the Hacktoberfest Open-Source AI Challenge.

🥾 TrailReady AI — Offline Trail & Route Planner

Submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

What I Built

TrailReady AI is an offline trail and route planner designed to minimize screen time so users can spend less time looking at devices and more time enjoying nature.

Built for hikers, runners, birdwatchers, and outdoor enthusiasts, the app takes a user's location, activity type, target distance, and preferred terrain, then generates a tailored route summary and packing checklist in seconds.

By running an open-weight AI model directly on the user's laptop, TrailReady AI requires zero cloud APIs or cellular signal—making it a reliable, privacy-first tool you can use even when deep in the woods or off the grid.

Demo

This app is designed to run locally.

• UI & Usage Screenshot:

Code

Mrunalikolte / trailready-ai

An offline, screen-free route planner & packing checklist generator powered by local Llama 3.2 open-weight AI

How I Built It

This project is built entirely on open-source AI tools and local execution frameworks:

• Local Inference via Ollama: Ollama serves as the local model harness, hosting and running open-weight LLMs directly on the machine's hardware with no external network calls.
• Open-Weight AI Model: Powered by Meta's Llama 3.2 (3B)—a compact, fast, and memory-efficient open-weight model optimized for edge devices.
• Streamlit Framework: Used to construct a simple, interactive Python web UI for capturing trail preferences.
• System Prompt Design: Uses structured system instructions to enforce strict output brevity, formatting the response into two clear sections kept under 150 words total.

Why Does Open Innovation Matter?

Open innovation is essential to the core purpose of this project:

1. 100% Offline Capability: When hiking or exploring trails without cellular reception, proprietary cloud APIs fail completely. Running an open-weight model locally ensures your planner works everywhere.
2. User Data Privacy: Your location, itineraries, and activity schedules stay on your device rather than being sent to third-party servers.
3. Zero Operating Cost ($0): Requires no API tokens, monthly subscriptions, or usage caps.
4. Full Model Independence: Open infrastructure allows swapping models (such as Gemma or Qwen) without vendor lock-in.

Prize Categories

• Main Track: Week 1 Challenge - Touch Grass
• Best Local Model Integration: Standard Ollama & Open-Weight Execution