The latest articles on DEV Community by Mrunali Kolte (@mrunalikolte). https://dev.to/mrunalikolte https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fthepracticaldev.s3.amazonaws.com%2Fi%2F99mvlsfu5tfj9m7ku25d.png https://dev.to/mrunalikolte en Mrunali Kolte Thu, 08 Oct 2026 20:19:17 +0000 https://dev.to/mrunalikolte/an-offline-trail-planner-built-with-llama-32-ollama-and-streamlit-for-the-hacktoberfest-24gn https://dev.to/mrunalikolte/an-offline-trail-planner-built-with-llama-32-ollama-and-streamlit-for-the-hacktoberfest-24gn <h1> 🥾 TrailReady AI — Offline Trail &amp; Route Planner </h1> <p><em>Submission for the <a href="https://dev.to/challenges/hacktoberfest-week1-2026-10-05">Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass</a></em></p> <h2> What I Built </h2> <p><strong>TrailReady AI</strong> 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. </p> <p>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. </p> <p>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.</p> <h2> Demo </h2> <p>This app is designed to run locally.</p> <ul> <li> <strong>UI &amp; Usage Screenshot:</strong> <img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fayelhsi95qn6c2bsc9o4.png" alt="TrailReady AI Interface" width="800" height="346"> </li> </ul> <h2> Code </h2> <div class="ltag-github-readme-tag"> <div class="readme-overview"> <h2> <img src="https://assets.dev.to/assets/github-logo-5a155e1f9a670af7944dd5e12375bc76ed542ea80224905ecaf878b9157cdefc.svg" alt="GitHub logo"> <a href="https://github.com/Mrunalikolte" rel="noopener noreferrer"> Mrunalikolte </a> / <a href="https://github.com/Mrunalikolte/trailready-ai" rel="noopener noreferrer"> trailready-ai </a> </h2> <h3> An offline, screen-free route planner &amp; packing checklist generator powered by local Llama 3.2 open-weight AI </h3> </div> </div> <h2> How I Built It </h2> <p>This project is built entirely on open-source AI tools and local execution frameworks:</p> <ul> <li> <strong>Local Inference via Ollama:</strong> Ollama serves as the local model harness, hosting and running open-weight LLMs directly on the machine's hardware with no external network calls.</li> <li> <strong>Open-Weight AI Model:</strong> Powered by Meta's <strong>Llama 3.2 (3B)</strong>—a compact, fast, and memory-efficient open-weight model optimized for edge devices.</li> <li> <strong>Streamlit Framework:</strong> Used to construct a simple, interactive Python web UI for capturing trail preferences.</li> <li> <strong>System Prompt Design:</strong> Uses structured system instructions to enforce strict output brevity, formatting the response into two clear sections kept under 150 words total.</li> </ul> <h2> Why Does Open Innovation Matter? </h2> <p>Open innovation is essential to the core purpose of this project:</p> <ol> <li> <strong>100% Offline Capability:</strong> When hiking or exploring trails without cellular reception, proprietary cloud APIs fail completely. Running an open-weight model locally ensures your planner works everywhere.</li> <li> <strong>User Data Privacy:</strong> Your location, itineraries, and activity schedules stay on your device rather than being sent to third-party servers.</li> <li> <strong>Zero Operating Cost ($0):</strong> Requires no API tokens, monthly subscriptions, or usage caps.</li> <li> <strong>Full Model Independence:</strong> Open infrastructure allows swapping models (such as Gemma or Qwen) without vendor lock-in.</li> </ol> <h2> Prize Categories </h2> <ul> <li> <strong>Main Track:</strong> Week 1 Challenge - Touch Grass</li> <li> <strong>Best Local Model Integration:</strong> Standard Ollama &amp; Open-Weight Execution</li> </ul> devchallenge hf26challenge Mrunali Kolte Tue, 06 Oct 2026 15:23:17 +0000 https://dev.to/mrunalikolte/hello-i-am-new-to-the-community-and-im-happy-to-join-1feh https://dev.to/mrunalikolte/hello-i-am-new-to-the-community-and-im-happy-to-join-1feh