This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What I Built
In modern software engineering, developers spend prolonged hours tethered to terminal output and high-luminance displays.
VerdantAI is an edge-native, open-source agronomy and ecological companion engineered around a single design objective: minimize digital interface duration and transition computational workflows into physical horticultural stewardship.
Key Capabilities:
- Hyperlocal Sowing & Frost Matrix: Computes frost probabilities, photoperiods, and succession schedules across temperate, subtropical, and monsoon agricultural zones entirely on-device.
- Pathology & Diagnostic Engine: Diagnoses foliar chlorosis, fungal pathogens (Erysiphales), and insect defoliation vectors with targeted organic remediation protocols.
- Hands-Free Speech Synthesis: Integrates the Web Speech API to provide audio remediation instructions in the field, allowing operators to tend garden beds without contaminating digital devices with soil.
- Field Protocols & Habitat Stewardship: A structured operational board that translates agronomic tasks (compost aeration, mulching, seed curation, avian transects) into logged field milestones.
-
Open-Weight Gemma 2 Integration: Delivers deep botanical reasoning powered by Google's Gemma 2 architecture, supporting local Ollama daemons (
gemma2:2b,gemma2:9b), OpenAI-compatible vLLM endpoints, or offline edge heuristics. - Procedural Environmental Sound Synthesizer: Leverages the Web Audio API to procedurally generate bioacoustics and atmospheric soundscapes (forest wind, rain, birdsong) in real-time with zero external media assets.
Demo
- Application Interface: Single-page architecture featuring a high-contrast dark botanical design system, deterministic calculations, and zero third-party tracking.
- Field Audio Integration: Audio narration allows hands-free inspection during outdoor physical tasks.
- Edge Diagnostic State: Real-time connectivity telemetry confirms the application operates fully disconnected from external networks.
Code
The complete source code is released under the MIT License and hosted on GitHub:
Kunal-CodeLab
/
verdant-ai
Edge-Native Open-Source Flora & Agronomy Companion for Hacktoberfest 2026
VerdantAI — Open-Source Offline Flora & Agronomy Companion
Decentralized botanical intelligence designed to minimize screen dependency and facilitate offline horticultural stewardship.
Developed for the Hacktoberfest 2026 Open-Source AI Challenge (Week 1: Touch Grass).
1. Executive Summary
Modern software development often leads to prolonged screen exposure and disconnection from physical ecosystems. VerdantAI is an edge-native, open-source agronomy platform engineered with a strict design constraint: minimize interface latency and compute duration to direct the user back to hands-on agricultural and ecological engagement.
Powered by Google's Gemma 2 open-weight model family and client-side botanical heuristics, VerdantAI provides localized agronomic guidance, specimen diagnostics, and field protocols without requiring continuous cloud connectivity or proprietary API tokens.
2. Core Architectural Pillars
2.1 Hyperlocal Sowing & Micro-Climate Engine
Calculates frost probability intervals, vernalization thresholds, and optimal sowing calendars across multiple global agricultural zones (Zone 5-11, South Asian plain agro-climates). All computations execute deterministically in the client…
Quick Start:
bash
git clone https://github.com/Kunal-CodeLab/verdant-ai.git
cd verdant-ai
# Open index.html directly in any modern browser. Zero build steps required.
Top comments (0)