Curie

Product Introduction

1. Definition: Curie is a specialized AI research assistant platform designed for rigorous scientific and academic literature review. It falls under the technical categories of AI-powered research tools, systematic review software, and document intelligence platforms.
2. Core Value Proposition: Curie exists to automate and validate the labor-intensive process of scientific literature review. Its core proposition is providing an AI assistant that prioritizes source verification, evidence-based analysis, and researcher-in-the-loop workflows for systematic reviews, meta-analyses, and structured data extraction, moving beyond the unreliability of generalist AI chatbots.

Main Features

1. Parallel Academic Search: Curie queries multiple, authoritative scientific databases—including PubMed, arXiv, Europe PMC, OpenAlex, and Semantic Scholar—simultaneously. This parallel search technology ensures comprehensive coverage and reduces the time researchers spend jumping between different search interfaces and managing disparate results.
2. Claim Verification with Source Attribution: This is Curie's signature feature. The AI doesn't just summarize; it extracts claims and then cross-references every statement directly against the original source text. The interface visually shows users what information is backed by a direct quote from the paper and what is an AI inference, dramatically increasing trust and auditability in the research process.
3. Structured Data Extraction & Project Management: Users can point Curie at a research document (like a PDF) to automatically extract key data (e.g., sample size, methodology, results) into a structured table or form. Crucially, each extracted value is linked to the supporting quote. Furthermore, Curie organizes this work into Projects and a Library, enabling management of full systematic review protocols, screening processes, PRISMA flow diagram generation, and collaborative oversight with manual approval gates.

Problems Solved

1. Pain Point: It solves the problem of information overload and verification fatigue in academic research. Manually screening hundreds of papers, extracting data, and verifying AI-generated summaries against sources is incredibly time-consuming and prone to human error.
2. Target Audience: Primary user personas include Academic Researchers, PhD Candidates, Medical Scientists, and R&D Teams conducting systematic literature reviews, scoping reviews, meta-analyses, or grant proposal research. It's also valuable for Research Librarians and Biotech/Pharma professionals involved in competitive intelligence and drug discovery.
3. Use Cases: Essential scenarios include: conducting a full PRISMA-compliant systematic review from protocol to synthesis; rapidly extracting structured data (e.g., chemical compounds, clinical outcomes) from a batch of PDFs for a meta-analysis; validating the factual accuracy of an AI-generated literature summary before including it in a paper; and maintaining an organized, citable library of research for a long-term project.

Unique Advantages

1. Differentiation: Unlike generalist AI chatbots (e.g., ChatGPT) or basic academic search engines, Curie is built specifically for the methodological rigor of science. It differentiates itself by embedding source verification into its core workflow and providing project management tools tailored for formal review processes, whereas competitors often lack this specialized focus and audit trail.
2. Key Innovation: The key technological innovation is its "evidence-first" AI architecture. Instead of generating a summary first and (maybe) citing sources later, Curie's process is engineered to anchor every piece of information to a retrievable source from the outset. This, combined with its multi-database parallel search and structured project scaffolding, creates a unique, end-to-end platform for validated research.

Frequently Asked Questions (FAQ)

1. How does Curie ensure the accuracy of its AI-generated summaries? Curie ensures accuracy through its core claim verification engine. It doesn't just summarize; it breaks down information into discrete claims and automatically retrieves the exact text passage from the source document that supports each claim, allowing the researcher to audit the AI's work instantly.
2. Can Curie help with creating a PRISMA flow diagram for my systematic review? Yes, Curie is designed to support the entire systematic review workflow. It can help manage the screening phases (title/abstract, full-text) and, based on your inclusion/exclusion decisions within the platform, generate the data needed to populate a standard PRISMA flow diagram, streamlining the reporting process.
3. What databases does Curie search and is it better than Google Scholar? Curie searches specialized academic databases in parallel, including PubMed, arXiv, Europe PMC, OpenAlex, and Semantic Scholar. This is fundamentally different from Google Scholar. Curie's approach provides more structured, database-specific searching and consolidates results from these vetted sources into one verified, project-managed interface, reducing noise and improving source quality.
4. Is Curie suitable for extracting data from clinical trial papers? Absolutely. Curie's structured data extraction feature is ideal for pulling specific data points from clinical trial PDFs, such as patient demographics, intervention details, primary outcomes, and statistical results. The critical advantage is that every extracted number or term is linked to a direct quote from the paper, ensuring traceability and reducing data entry errors for evidence synthesis.