Janela: A Local Gemma-Powered App for Outdoor Planning

Janela is a new open-source application designed to help users find the optimal time for outdoor activities like running, cycling, or gardening. Built for the Hacktoberfest 2026 'Touch Grass' challenge, the tool uses Java-based logic to process weather data from Open-Meteo, ensuring accurate calculations for temperature, UV index, and rain probability. To provide a personalized experience, the app integrates a local instance of Google's Gemma 3 (4B) model via Ollama. The developer emphasizes a 'deterministic-first' architecture where the AI is strictly limited to generating natural language explanations, while the core decision-making remains in robust, unit-tested Java code. This approach prevents common LLM hallucinations regarding arithmetic or scheduling. Janela runs entirely locally, requires no API keys, and offers a privacy-focused alternative to cloud-based weather assistants, demonstrating how small, efficient models can effectively enhance user-facing applications without compromising data privacy or reliability.
This is a summary. Read the full article at the original source:
Dev.toRelated stories
Developing a Fuel Availability Monitor Using AI: A Product Manager's Experience
Igor, a product manager at Outlines Tech, shared his experience building a custom tool to track the availability of AI-95 gasoline at gas stations. Fr…
Microsoft Brings MAI Code 1.1 Flash Local Inference to GitHub Copilot
Microsoft has announced the integration of MAI Code 1.1 Flash into GitHub Copilot, enabling local inference for coding workflows. This update allows d…
The author shares their experience using Claude Code in a native Windows environment, bypassing WSL. The main challenges when working with an AI agent…



