Open core Claude-Code for Bioinformatics (github.com)

Lobster AI is an open‑source, “open‑core” bioinformatics agent platform that lets researchers describe analyses in natural language and get end‑to‑end multi‑omics workflows (starting with transcriptomics) without coding. From a simple CLI ("lobster chat" or "lobster chat --reasoning") it will download GEO datasets, run single‑cell QC, normalization, clustering and UMAPs, annotate cell types, identify marker genes, run pseudobulk and differential expression (pyDESeq2), and produce publication‑ready visualizations and reproducible workspaces. It also supports common input formats (CSV, Excel, H5AD, 10X), batch correction, complex experimental designs, literature mining (PubMed), proteomics pipelines (DDA/DIA, Olink), and multi‑modal integration. Technically, Lobster integrates specialized AI agents with existing bioinformatics tools and scalable cloud execution; it requires Python 3.12 and an LLM API key (Anthropic Claude recommended, or AWS Bedrock). The repo (github.com/the-omics-os/lobster-local) installs via Make and is Apache‑2.0 licensed (docs CC‑BY‑4.0). For the AI/ML community this lowers the barrier to sophisticated, reproducible analyses by combining LLM orchestration with domain tools, accelerating prototyping and teaching—while raising important considerations around LLM cost, data privacy/provenance, and potential hallucinations when interpreting results produced via external LLM providers.