Tendril: a garden assistant you never have to open
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass . What I Built My balcony has a little garden: two Indian hibiscus, a money plant, jasmine, a blue aprajita, mint, tulsi and sadabahaar. All in pots. I water them every morning and every evening, without fail. But a routine doesn't look at the sky. It waters the plants an hour before a downpour. It doesn't…
This is a story about Tendril, a garden assistant that you never have to open. The author built Tendril to help maintain their balcony garden, which includes plants such as Indian hibiscus, money plant, jasmine, blue aprajita, mint, tulsi, and sadabahaar. The assistant plans the week's gardening tasks based on the weather forecast, ensuring that the plants receive the right amount of care.
Tendril plans the tasks for the week every Sunday at 7 AM, generating at most two simple tasks per day, each described in under six words. It checks the sky every morning and rewrites the tasks if rain is predicted or if there is a heat spike. The plan is presented as a single picture, designed to fit on a phone's lock screen with the top third left empty for the clock.
The assistant can be customized to record what actions were performed, either by tapping "Done" or by speaking a note using voice commands. It supports both English and Hindi, and can be set up by entering the city name, adding plants, and scanning a QR code. Once installed, it can be added to the home screen.
Tendril's core functionality is built using Python, FastAPI, SQLite, Qwen 2.5 (3B) open-weight model, Ollama, faster-whisper speech model, Open-Meteo for weather forecast and city search, and Pillow for drawing the plan picture. The entire setup can be run with a single Docker Compose command.
The author emphasizes that Tendril is designed to be minimalistic and efficient, using a small open model that runs on the user's own machine to minimize costs and ensure privacy. The system checks the model's output using Pydantic, and if necessary, falls back to a rule-based planner to ensure proper task planning. The assistant is set up for voice notes and can catch common mistakes, such as forgetting to water plants.
The final product is a calm, unobtrusive garden assistant that helps maintain the plants without the need to constantly check a phone or app. The author describes their experience using Tendril on their balcony garden, noting that the assistant successfully adjusted the watering schedule in response to a rainy week.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
