@xiaohu: Anthropic launches Claude Science, an AI workbench for scientists with over 60 research skills built in. It is an application installed on your own computer or server: you ask an AI scientific questions in plain language, and it mobilizes dozens of specialized tools to query data, run analyses, draw charts, and draft manuscripts…

X AI KOLs Timeline Products

Summary

Anthropic has launched Claude Science, an AI workbench for scientists with over 60 built-in research skills. It supports local deployment and HPC clusters, and can autonomously draft computing tasks and review results.

Anthropic launches Claude Science An AI workbench for scientists, with over 60 built-in research skills It is an application installed on your own computer or server: you ask an AI scientific questions in plain language, and it mobilizes dozens of specialized tools to query data, run analyses, draw charts, and draft manuscripts. Every step of the output can be traced back to how it was produced. You can use it locally (macOS/Linux) like Jupyter Notebook, or on a remote machine via SSH or an HPC login node. → The application includes over 60 pre-configured skills and connectors covering genomics, single-cell analysis, proteomics, structural biology, and cheminformatics, backed by hundreds of specialized data sources (UniProt, PDB, Ensembl, etc.) as well as journals and preprint resources. → It can autonomously draft computing tasks, and after obtaining user consent, submit them to the user's own HPC cluster or Modal cloud GPU, scaling analysis from a single GPU to hundreds, while the original data always remains in the user's own system. → It includes a built-in reviewer agent that constantly checks whether citations in the generated content are real, whether numbers match the computational process, and whether charts are consistent with the code that produced them. If issues are found, it automatically corrects them.
Original Article
View Cached Full Text

Cached at: 07/01/26, 08:02 AM

Anthropic Releases Claude Science

An AI workstation for scientists, with over 60 built-in research skills

It’s an app that runs on your own computer or server: you ask a scientific question in plain language, and it orchestrates dozens of specialized tools to look up data, run analyses, generate charts, and draft manuscripts—every step along the way is fully traceable.

You can use it locally on macOS/Linux like you would Jupyter Notebook, or remotely via SSH or an HPC login node.

→ The app includes over 60 preconfigured skills and connectors covering genomics, single-cell biology, proteomics, structural biology, and cheminformatics, connected to hundreds of specialized data sources (UniProt, PDB, Ensembl, etc.) as well as journals and preprint repositories.

→ It can autonomously draft computational tasks, submit them to your own HPC cluster or Modal cloud GPUs after receiving your approval, scaling from a single GPU to hundreds, while keeping your raw data within your own systems at all times.

→ A built-in review agent continuously checks the generated content: verifying that citations are real, numbers match the computation, and charts are consistent with the code that produced them—automatically fixing any issues it finds.

Similar Articles

@grapeot: Why Anthropic's newly released Claude Science, seemingly lacking the 'grand narrative' of scientific research, actually hits the real pain points of scientists' daily work? Many people expect an 'AI scientist' that can reason autonomously and guide the way. But in reality, the bottlenecks in life sciences and pharmaceuticals are often a series of extremely...

X AI KOLs Timeline

Anthropic's Claude Science desktop application focuses on solving the digital grunt work in scientists' daily research, such as data integration and supercomputer scheduling, rather than pursuing a grand-narrative AI scientist. It lowers the barrier to scientific computing through a natural language interface.

@sodawhite_dev: https://x.com/sodawhite_dev/status/2067413032544940062

X AI KOLs Timeline

The article analyzes Anthropic's 400,000-session report on Claude Code, pointing out that AI programming tools are changing the division of labor between humans and AI. Domain knowledge is more important than coding ability. Expert users can enable AI to perform more complex tasks, while verification and task decomposition capabilities become core competitive advantages.