MedAI Tech: Using 40 Million Documents + AI Agents to Build China's OpenEvidence

Over the past year, OpenEvidence, a U.S.-based medical AI agent, has continued to gain momentum in the capital market: it completed four rounds of financing within a year, with the latest valuation reaching $12 billion.

The core of its success lies in avoiding high-risk diagnosis and treatment processes, focusing on answering doctors' knowledge-related questions, and using AI to improve the efficiency of evidence-based medicine and scientific research. Currently, its products have covered more than 40% of practicing physicians in the United States.

This case clearly outlines an emerging track: centered on "AI + Evidence-Based Medicine," it reconstructs the way doctors access and understand medical knowledge, and unleashes a vast market space through efficiency improvements.

Turning our attention to China, this trend is resonating in sync with clear policy guidance. The National Health Commission has proposed to "deepen the intelligent application of medical scientific research," promoting the use of AI across all research processes, including literature review and data analysis. Against this backdrop, exploration around medical Q&A and knowledge-assisted tools is becoming a significant direction for AI healthcare in China.

It is in this context that MedArk Knowledge (Beijing) Technology Co., Ltd. (hereinafter referred to as "MedArk Technology") launched Info X Med, attempting to"Medical Literature Database + AI Agent"As the technical foundation,Avoid the high uncertainty of diagnostic AI, and start from more fundamental and universally applicable knowledge links., explore a localized path that is closer to the actual workflow of Chinese doctors.

The rigor of Info X Med stems from the deep empowerment of a tens-of-millions-scale medical database, forming full-chain quality control from data source to content output.

Info X Med Platform Model Diagram Source: Yifu Technology

Authoritative, Comprehensive, and Timely Literature Base

The platform integrates over 40 million global medical literature, international guidelines, and funding data, covering resources from top global publishers such as Elsevier and Wiley. The full-text PDF inclusion rate is as high as 95%, ensuring the comprehensiveness and authority of the literature foundation from the source.

It not only includes clinical research and Meta-analysis but also focuses more on integrating localized research data targeting the Chinese population. Through a real-time update mechanism, the platform can quickly synchronize cutting-edge achievements from top academic conferences such as the American Academy of Orthopaedic Surgeons (AAOS) and the Orthopaedic Research Society (ORS), making the literature system both timely and clinically relevant.

Technology-Driven, Clear and Traceable Chain of Evidence

The platform, based on its proprietary medical database and trained professional models, has achieved a 96.7% accuracy rate in medical semantic understanding. When assisting in writing, the AI can automatically annotate each conclusion with a reference number and allows one-click tracing back to the original document's digital identifier (DOI/PMID), ensuring all viewpoints are verifiable.

During analysis, the system can intelligently link key points from the latest literature with relevant clinical research data, forming a complete logical chain from "literature recommendation" to "research validation" and then to "practical application," effectively supporting rigorous clinical and scientific argumentation.

Proven Professional Reliability

In the practical application in clinical departments of several well-known Class III, Grade A hospitals, the platform has shown significant results: the average time doctors spend on extensive literature reading is reduced by 80%, with no omission of important research progress. The generated content framework meets the standard requirements of domestic core journals such as the Chinese Journal of Medicine.

With the convenience of accessing literature (no internet restrictions, timely updates) and the highly rigorous generation (rooted in a vast amount of authoritative literature), the platform effectively avoids "academic hallucinations" and significantly surpasses general-purpose AI tools in professional reliability.

Building on this solid foundation of literature, Info X Med deeply integrates intelligent search with advanced literature processing functions, aiming to upgrade the experience of literature retrieval and reading from "searching" to "immediate understanding and application." This includes:

● Intelligent Search and Instant Presentation: After doctors search on the platform, the results are not only accurate but also allow direct access to the original text. For foreign language literature, the system provides deep AI instant professional translation in the original format, supporting quick conversion of the full text or paragraphs, helping doctors overcome language barriers and achieve instant browsing.

● Efficient Content Extraction and Organization: Faced with valuable literature, doctors can export core charts and data from articles with one click, making it convenient to directly use them in reports or for study. Going a step further, the platform supports dialogue with PDF documents. Doctors can ask questions in natural language without having to read the entire text, allowing AI to quickly analyze research design, core findings, data differences, etc., guiding doctors in deep thinking and selective reading.

● Knowledge Accumulation and Efficient Output: Based on the intelligent interpretation of literature, the system can automatically summarize key points and further generate a well-structured and accurate draft of a PPT file. This greatly simplifies the preparation of materials for doctors conducting academic presentations or teaching lectures, significantly shortening the chain from knowledge acquisition to output.

Through the above functions, Info X Med transforms traditional literature databases into an intelligent knowledge learning and working platform that integrates search, reading, analysis, organization, and output, directly addressing the pain point of doctors' lack of time and turning massive literature from a burden into an efficiently usable asset.

Building on a solid foundation of literature, Info X Med, developed by MedTech, has not stopped at the evolution of search tools but has proactively constructed a dual-engine architecture comprising a "medical database + self-trained medical AI large model."

This approach avoids the potential incompatibility issues that may arise from directly applying AI tools developed in foreign English-language contexts and based on a single evidence system. Instead, it focuses on enhancing the work efficiency of Chinese doctors in complex scenarios of "medical practice, teaching, and research" through localized, scenario-specific medical AI agents.

The platform clearly positions AI as a "productivity-enhancing assistant" for doctors, with the core value of empowering rather than replacing them. Based on this, three core intelligent agents have been developed:

Research Intelligence: Accelerating the Entire Process from Literature Digestion to Initial Draft of Papers

Research Intelligence Agents aim to revolutionize how doctors interact with literature, accelerating the entire process from knowledge digestion to outcome generation. Doctors can directly upload or select PDF articles, then engage in natural language conversations with AI, instructing it to automatically extract key elements such as study design, core data, and conclusion disputes. This compresses the traditionally time-consuming task of intensive article reading from several hours into just a few minutes.

Moreover, when researchers input a broad field, the intelligent agent can analyze a vast amount of literature, organize existing research threads, and intelligently suggest potential directions that have not been fully explored, providing high-value insights for project initiation.

During the output phase, intelligent agents trained on authoritative medical databases of the platform can assist in generating a complete framework for literature reviews or initial drafts of papers based on user instructions. They can also generate full-text files directly downloadable and editable according to professional SCI templates. The system can dynamically link to the latest research data, automatically update key arguments, and form a rigorous logical closed loop from literature recommendation, research validation to practical application, significantly improving the overall efficiency of clinical research manuscript preparation.

Evidence-based Intelligent Agent: A 24-hour Clinical Assistant Based on Natural Language Q&A

This is the core intelligence directly serving the clinical front line. It deeply integrates seven authoritative databases: clinical pathways, drug instructions, disease knowledge bases, clinical guidelines, systematic reviews, Meta-analysis, and evidence-based data, addressing the issues of ambiguity, hallucination, and inefficiency in traditional keyword searches.

Doctors only need to input specific clinical questions, and the system can precisely filter out relevant latest guidelines and clinical studies from a vast amount of evidence within one minute. It not only extracts core conclusions and levels of evidence but also generates structured clinical decision-making references with original source citations.

This Q&A-style interaction simulates the expert teaching and research process, enabling doctors who have just entered clinical work to quickly obtain evidence-based decision-making references in time-pressured scenarios such as on-call duties and consultations, effectively boosting the confidence and efficiency of basic clinical decision-making for frontline doctors.

Library Intelligence: A Personalized Learning Engine Covering All Fields of Medicine

The Library Intelligent Agent focuses on lifelong learning and capability enhancement for doctors, serving as a personalized learning engine that covers all medical fields. It contains over 1.2 million questions, comprehensively encompassing the knowledge system required for clinical practice and research.

Doctors can quickly enter an efficient practice and consolidation phase based on their own discipline, professional title, or specific exam preparation goals. The system automatically tracks and cumulatively records incorrect questions, intelligently generating a map of individual weak areas to assist doctors in targeted review. Meanwhile, the platform promises a pure, ad-free, and hidden-cost-free learning environment, dedicated to creating a truly focused and efficient knowledge consolidation space for doctors.

To ensure the quality of all content reading and output, Info X Med has built a rigorous quality control system, including professional language translation with an accuracy rate of up to 96.7%, advanced search formulas surpassing PubMed, self-trained professional medical large models, and other technologies. All generated viewpoints are linked with literature for verification. Currently, its medical content generation accuracy rate exceeds 95%, with a hallucination rate below 4%, achieving a leading balance of efficiency and precision among market-oriented products in the vertical field.

Through the main function of literature retrieval + three major intelligent agents, Info X Med precisely injects the capabilities of its self-trained large medical AI model into three core scenarios: scientific research exploration, clinical decision-making, and lifelong learning. This truly realizes the dual-engine synergy of "medical database" and "professional AI agent," becoming a powerful tool to significantly enhance the comprehensive efficiency of doctors in China.

The "Knowledge Dilemma" of Doctors: A Definite, Nationwide Market Opportunity. Research data shows that the average weekly online study time for the group of doctors in China is 14.1 hours, which includes intensive literature review.

At the same time, Frost & Sullivan predicts that the scale of China's pharmaceutical digital marketing market will grow to 356.8 billion yuan by 2030. These two figures point to the same hub: the process by which doctors efficiently acquire compliant knowledge is becoming a core node in the future transmission of medical value.

In recent years, the development path of medical AI has gradually diverged. On one end are applications directly involved in diagnosis and treatment, such as AI imaging diagnostics, surgical navigation, and clinical decision support systems. These products intervene directly in clinical settings, with high regulatory thresholds, long R&D cycles, and uncertain commercialization processes. On the other hand are efficiency tools that empower doctors' knowledge workflows, such as literature search, clinical Q&A, research assistance, and evidence-based analysis platforms. These tools have controllable risks and rigid demand, leading to their rapid rise.

Info X Med, launched by MedAffinity, is closer to the OpenEvidence model and optimized, cutting in from the source of large intelligent models, with particular emphasis on the accuracy of clinical Q&A and the integrity of the evidence-based chain.

Currently, MedTech's business maintains healthy growth. In the eyes of the capital market, its value is not only derived from its current financial performance but also deeply rooted in the long-term scalability of its industry sector. As a medical knowledge platform, its user base and usage frequency have yet to reach their ceiling. With a solid entry point, it possesses a natural platform network effect that extends into multiple scenarios such as research, clinical practice, education, and industry compliance. This unique ecosystem scalability forms the core logic supporting its long-term value expectations.

In the future, with the deepening of China's medical consortium construction and the standardized popularization of clinical research, the vision of VCBeat (MedTech) is becoming increasingly clear——By continuously building this robust knowledge infrastructure, precious time and decision-making initiative will truly be handed back to Chinese doctors.