Blog

Smaller models hallucinate. So do expensive ones. Ground any LLM with the Scite MCP for accurate citations at a fraction of the cost.

May was about giving you more room to work and more ways to organize what you find. We introduced a new plan for power users, brought a better way to organize your research into the platform, taught Assistant to read your own documents, and expanded the data you can reach through Scite.

Scite MCP now reaches 719K+ clinical trials and 3.5M+ grants alongside Smart Citations, so papers, trials, and funding sit within a single prompt.

Collections brings your research into one place — a dedicated space to gather papers, monitor them over time, and run focused, scoped questions against exactly the set you care about.

April was about expanding where Scite's intelligence lives and what it can search. We extended the Scite MCP beyond the published literature to cover patents, clinical trials, and grants; launched a new microsite to showcase what Scite + AI clients can produce; and made it faster than ever to get started with one-tap Google sign-in.

Search 90M+ patent families alongside academic literature in Scite. Patent Search and Scite Assistant now unify IP and research data in one place.

Scite is now available as a verified connector in the Claude directory. Add it in one click to ground Claude's answers in real peer-reviewed literature with citations you can verify.

Students can connect Claude, ChatGPT, and other AI tools to Scite via MCP to search literature, check citation context, and access full texts. But their institutional access comes from their Scite account, not their AI tool — so the way patrons set up the connection matters. Here's what librarians need to know to get students linked to their library's license the first time.

This month we focused on making Scite Assistant more flexible and easier to use — from summarizing articles by DOI, to saving your preferred settings, to dragging citations wherever you need them. We also refreshed our website with new landing pages and a redesigned blog.

Claude Code with the Scite MCP changes the math on dissertation research pretty significantly. You get an AI assistant in your terminal that can actually search scientific literature and tell you not just what's been published but how it's been cited since.

Starting in 1977, Eugene Garfield's Citation Classics series at ISI invited the authors of highly cited papers to write short commentaries on their own work. Over 15 years, more than 4,000 scientists told the stories behind their most influential publications: the failed experiments, the lucky breaks, the collaborators who made the difference. It was one of the most human things ever done with citation data.

Admin Analytics gives administrators at universities, enterprises, and research institutions full visibility into how their researchers are using the Scite MCP inside their AI tools — what content is being accessed, how it's being used, and who's using it. And it connects that usage data directly to your collections infrastructure so you can act on it.

February was about making Scite more powerful, more connected, and easier to use. We expanded how you can access Scite's citation intelligence, removed barriers to deeper research, and made significant improvements across Assistant, Search, and Dashboards.

Most AI tools have a fundamental problem: they aren't grounded in real evidence. That's the problem Model Context Protocol (MCP) solves.

Scite is now available as a verified plugin in the ChatGPT plugin directory. Add it in one click to ground ChatGPT's answers in real peer-reviewed literature with citations you can verify.

Starting today, you can connect Scite to the AI tools you already use. Search 250M+ articles, evaluate research with Smart Citations, and access full-text PDFs - all without leaving your workflow.

We recently asked researchers, information professionals, and academics a simple question: Do you trust LLM answers without citations? 88% said no. When we asked what matters most when using generative AI, accuracy and trustworthiness topped the list, followed closely by citation quality and source verification.

January focused on expanding what researchers can explore, improving how work gets organized, and strengthening the foundation for faster innovation throughout 2026.

December brought a set of meaningful improvements to Scite Assistant focused on clarity, usability, and deeper exploration.