Google Gemini
Visit official siteWhat It Does
Multimodal AI model from Google capable of serving as a technical architect — translating plain-English requirements into detailed implementation instructions, generating structured planning documents, roleplaying user scenarios for testing, and producing rich content. Available in Pro (capable) and Flash (fast/lightweight) tiers.
How We Use It
Deployed as the technical architect layer, translating plain-English product requirements into detailed step-by-step implementation instructions for the OpenClaw coding agent. Also used to generate fictitious student test profiles, roleplay student interactions during testing sessions, and produce session-reset manifestos. Chosen over direct coding-agent communication to reduce token consumption on expensive API calls.