Core Sources
Start with the essentials
PubMed
Biomedical and life sciences literature database from the U.S. National Library of Medicine.
FundingNIH
Policy, institutes, grants, and research updates from the U.S. National Institutes of Health.
DiscoveryConnected Papers
Explore paper relationships visually to find prior work, follow-up studies, and adjacent ideas.
DiscoveryResearchRabbit
Build citation maps, follow connected papers and authors, and track literature discovery visually.
Useful Additions
Common tools researchers reach for
Google Scholar
Broad citation search across journals, theses, books, and preprints.
Open AccessPubMed Central
Free full-text archive of biomedical and life sciences journal literature.
PreprintsbioRxiv
Preprint server for biology that is often useful for fast-moving literature tracking.
PreprintsmedRxiv
Preprint server for health sciences and clinical research updates.
DiscoverySemantic Scholar
AI-assisted academic search with citation graphs, paper summaries, and related work.
IdentityORCID
Persistent researcher identity used across journal submission, grant, and affiliation systems.
QuestionnairePsychology Scales
English originals and verified Chinese sources for common psychology questionnaires.
AI Tools
Log in, route, and get started
400Cloud
Log in to 400Cloud, then connect your AI coding tools through CC Switch.
AppCC Switch
The official home for the open-source AI coding CLI manager.
StartGetting Started
Quick start, installation, and first-run setup from the official docs.
ProvidersProviders
Add providers, switch between them, and review usage details.
TutorialsTutorials
Official and community guides for getting started, integrations, and troubleshooting.
RoutingLocal Routing
Route Codex into Claude Code with CC Switch's local routing flow.
More official reads: DeepSeek routing, auth preservation, and the release notes.
Skills
Mounted at the same site
The skills directory now lives at researchtoolkit.magicmightywhales.top/skills. It reuses the same server and site shell, so you do not need a new machine just to start.
For now, the page is the front door. Later we can attach the real skills runtime behind the same path and split infrastructure only if the workload grows.