Anthropic is rolling out Claude Science as a beta desktop app that folds everyday research tools into a single AI session, generating artifacts that carry traceable code, environment details, and plain-language notes so experiments stay reproducible from first sketch to final manuscript.
Introducing Claude Science, a new app designed with every stage of research in mind.
Artifacts traced to their code, environments managed on demand, and 60+ optional scientific databases that you can connect.
Available now in beta.
Pro, Max, Team, and Enterprise users on macOS or Linux can enable the workbench right away, with a special discounted Team plan offered to academic labs and nonprofits; free-tier access and Windows support remain unspecified for now.
Specialist agents connect to more than 60 scientific databases while compute runs locally or on user-chosen resources, keeping sensitive data in place and letting labs add their own packages without handing everything to the cloud.
Positive users praise Anthropic's Claude Science Beta app for research workflows and its database integrations, while negative users accuse the company of fraudulent billing and charging for free tools.
No Digg Deeper questions have been answered for this story yet.
@claudeai Do i have to be a scientist to use this? 👀
@claudeai You can ask Claude Science to install LigandAI (pip install ligandai) and reference http://ligandai.com/sdk to generate nearly instant peptides and come up with target discovery :)
@claudeai
You can ask Claude Science to install LigandAI (pip install ligandai) and reference http://ligandai.com/sdk to generate nearly instant peptides and come up with target discovery :)
Introducing Claude Science, a new app designed with every stage of research in mind.
Artifacts traced to their code, environments managed on demand, and 60+ optional scientific databases that you can connect.
Available now in beta.
@claudeai Great to see more tools for AI + science! For anyone looking for an open-source, model-agnostic option, we’ve been building K-dense BYOK: https://github.com/K-Dense-AI/k-dense-byok
@claudeai @_LaurentB Connecting scientific databases to predict the World Cup winner:
@claudeai The datasets need a long-term home too, somewhere verifiable that won’t disappear when a lab changes tools or a grant runs out.
Verifiable storage earns its keep here.
@claudeai Scientific research assistants right now
@claudeai Your AI can sandbox and create whatever adaptation of a fake BS reference and bias to inject and manufacture social change.
Layer lies on lies on lies.
Systems so fraudulently complex they can’t be verified or denied.
If you had an agenda…I guess you could use it for that.
@claudeai
Biology is becoming software. Whoever becomes its operating system owns the next decade.
This is the Claude Code playbook again. Claude Code didn't win by being the best coder. It won by becoming where programmers work. The model underneath went quiet and the workspace ate everything.
Now they're pointing that at biology. Biology is just the beachhead. Law, finance, engineering are the same shape, a profession drowning in tools, waiting for one environment to swallow them.
Introducing Claude Science, a new app designed with every stage of research in mind.
Artifacts traced to their code, environments managed on demand, and 60+ optional scientific databases that you can connect.
Available now in beta.
@claudeai Me: Where's that meme again
@claudeai
@claudeai
@claudeai Does this surreptitiously alter user's work deemed a threat vector as well? What possible scientific use can you truly offer when you erode the user base trust in the model? Rollout central doesn't address what Dario set in motion or how anyone could trust your ai. Imo.
@claudeai
@claudeai everyone reads this as another chatbot. the real lever is the 60+ databases you can connect — the science bottleneck was always getting legal access to the data, not reasoning over it. jumper spent a decade on exactly that. the plumbing is the product.
@claudeai
Vibe peptides! ;)
You can ask Claude Science to install LigandAI (pip install ligandai) and reference http://ligandai.com/sdk to generate nearly instant peptides and come up with target discovery :)
@claudeai
