How Pollinator Patch Uses AI
A plain-language account of where AI is used in this app and this site, and where it is not.
What AI does here
Ask Patch answers questions using large language models from Anthropic. Patch Vision generates yard renderings with third-party image models (Google Gemini and models hosted at fal.ai). Patch ID identifies plants from photos using Anthropic's vision models, with a Pl@ntNet fallback.
What AI does not do
AI does not choose which plants the app recommends. Plant recommendations come from a deterministic engine built on ecoregion, sun, and HOA strictness data. Rebate amounts, statutes, and toxicity claims are cited to their official sources, not generated.
Writing on this site
Much of the writing on this site, including blog posts and plant care answers, starts from our own research, field notes, and recorded voice memos, and is drafted with AI assistance. Every piece is reviewed and edited before publishing, and Stephen Janacek, the founder, holds editorial responsibility for everything published here. Claims about plants, rebates, and laws are checked against the cited sources.
Images on this site
Patch Vision before-and-after images are AI-generated renderings of real yard plans. Most blog illustrations are also AI-generated. Photos of individual plant species in the app and in plan guides are real photographs, not renderings.
Watermarks
AI providers now embed invisible watermarks and signed metadata in generated text and images, including the drafting tools used here. Content on this site may carry these marks. We do not attempt to remove them. A watermark indicates that a tool was used in producing the content, not who authored it; authorship and editorial responsibility are described above.
Questions or problems
Every Patch Vision rendering, Patch ID result, and Ask Patch answer in the app has a report link if something looks wrong. For anything else, reach out directly.
See also the Privacy Policy and Terms of Service.
Last updated: August 2026