PlantLab - structured visual diagnosis to complement HAGR's environmental monitoring

[plantlab-ai]

Hey - been looking through HAGR and saw you're using the OpenAI API for environmental text summaries.

I built something that covers the visual side: what the plant actually looks like. PlantLab is a Home Assistant integration that diagnoses plant health from camera photos. 30 specific conditions and pests, confidence scores, all returned as structured JSON your automations can branch on. Purpose-built vision models trained specifically on cannabis, not an LLM wrapper.

The practical difference from LLM-based analysis is the output format. You get {"class_id": "magnesium_deficiency", confidence": 0.94} instead of a paragraph. Your automation can act on that directly - adjust dosing, dim lights, trigger an alert - without parsing natural language.

GitHub: https://github.com/plantlab-ai/home-assistant-plantlab
Free tier: 3 diagnoses/day

HAGR tracks the inputs (temp, humidity, VPD). PlantLab tracks the plant's response. Different sensor modality, same grow. Let me know if you see a fit. Cheers!