The Gardening Assistant
An AI-powered open-source gardening planning tool that uses local inference via Google Gemma. Based on a user's sunlight conditions, garden type, experience level, and available space, it recommends plants and generates a personalized four-week outdoor task plan.
Timeline (last 90 days)
The Gardening Assistant (TGA) is an AI-driven gardening planning tool that recommends suitable plants based on the user's sunlight conditions, garden type, experience level, and available space.
TGA was built by developer DaKaufeeBoii for the "Touch Grass" challenge at Hacktoberfest.
TGA generates personalized gardening plans that include three targeted plant recommendations, care instructions, and a four-week outdoor task checklist.
TGA's weekly checklist recommends that users spend approximately 15 to 25 minutes outdoors each session.
TGA supports models of varying scales such as gemma:2b and gemma2:9b, and can be flexibly configured based on hardware conditions.
TGA employs a rule-based deterministic recommendation engine as a fallback, enabling gardening plan generation even when the local AI model is unavailable.
When calling Gemma, TGA injects relevant facts from a curated JSON plant dataset into the prompt, grounding the model's output in verified plant information to reduce hallucinations.
TGA uses a lightweight implementation that directly concatenates filtered plant data into the context, omitting infrastructure such as vector databases, in an approach similar in principle to RAG.
Local inference allows TGA to deliver personalized recommendations without sending garden profiles to a remote model provider, thereby protecting privacy.
The author has planned extensions such as voice narration (ElevenLabs), real-time frost date lookup (SerpApi), and backend observability (Sentry Agent Tracing), but these remain planned extensions rather than completed features.
2 more timeline events
All Facts (12)
TGA's source code is hosted on GitHub at github.com/DaKaufeeBoii/The-Gardening-Assistant.
50%UnverifiedTGA was built by developer DaKaufeeBoii for the "Touch Grass" challenge at Hacktoberfest.
50%UnverifiedTGA generates personalized gardening plans that include three targeted plant recommendations, care instructions, and a four-week outdoor task checklist.
50%UnverifiedTGA's weekly checklist recommends that users spend approximately 15 to 25 minutes outdoors each session.
50%UnverifiedTGA supports models of varying scales such as gemma:2b and gemma2:9b, and can be flexibly configured based on hardware conditions.
50%UnverifiedTGA employs a rule-based deterministic recommendation engine as a fallback, enabling gardening plan generation even when the local AI model is unavailable.
50%UnverifiedWhen calling Gemma, TGA injects relevant facts from a curated JSON plant dataset into the prompt, grounding the model's output in verified plant information to reduce hallucinations.
50%UnverifiedTGA uses a lightweight implementation that directly concatenates filtered plant data into the context, omitting infrastructure such as vector databases, in an approach similar in principle to RAG.
50%UnverifiedLocal inference allows TGA to deliver personalized recommendations without sending garden profiles to a remote model provider, thereby protecting privacy.
50%UnverifiedThe Gardening Assistant (TGA) is an AI-driven gardening planning tool that recommends suitable plants based on the user's sunlight conditions, garden type, experience level, and available space.
50%UnverifiedTGA deliberately keeps the model's responsibilities focused rather than having an autonomous agent control the entire workflow, making the recommendation process easier to reason about and debug.
50%UnverifiedThe author has planned extensions such as voice narration (ElevenLabs), real-time frost date lookup (SerpApi), and backend observability (Sentry Agent Tracing), but these remain planned extensions rather than completed features.
50%