The arXiv paper details two hybrid Local Body, Remote Brain systems using Python orchestrators on Google Colab to call LLM backends for scientific tasks. DeepTS/DeepCollector performs large-scale time-series dataset curation, extraction, and deduplication while DeepScribe converts dense physics lecture visuals into structured reports via cellular RAG and distributed controls. The work also generalizes DeepTS toward knowledge graphs and applies the approach to high-energy physics under DeepQCD. Primary evidence is limited to system descriptions without quantitative benchmarks against prior agent baselines such as those in arXiv:2210.03629. Related patterns appear in autonomous lab systems reported in Nature 2023 and 2024 papers on self-driving laboratories.