XtalPi launches epiXora™, a life sciences–dedicated AI agent, accelerating the completion of its Multi-Agent R&D closed loop with a trusted decision-making foundation

XtalPi (2228.HK) today announced that itsThe professional-grade AI research agent epiXora™ has been fully upgraded and is now live, with registration for trial access open.As a critical front-end extension of XtalPi’s 'AI + Robotics + Multi-Agent' technology stack, epiXora™ deeply embeds authoritative domain-specific databases and R&D logic into its workflow, delivering trustworthy, traceable, and reusable decision support for core scenarios such as target selection, pipeline evaluation, and external partnership assessments. As the company's multi-agent matrix continues to mature,XtalPi is progressively deploying proprietary Scientific Agents across more research scenarios, establishing a digital intelligent hub for the industry to tackle complex innovation challenges.

In recent years, general-purpose AI tools have significantly improved information retrieval efficiency. However, in highly specialized fields such as life sciences, they often suffer from poor adaptability. In real-world scenarios like target assessment or R&D project initiation, general-purpose AI frequently falls short due to insufficient coverage of domain-specific data, misalignment between analytical logic and decision pathways, and a lack of rigorous evidence chains—creating a 'trust gap' that renders their research outputs unsuitable as direct inputs for decision-making. The industry urgently needs to evolve from 'general information retrieval' toward specialized Agentic AI. epiXora™ was designed precisely to address this industrial need. It not only retains the speed and breadth of general-purpose AI in processing vast volumes of information but also understands the highly rigorous decision logic inherent in biopharma contexts, directly addressing the industry’s core demand for certainty and credibility in conclusions—thereby elevating AI-powered research from basic information summarization to a reliable 'decision-support engine.'

To truly free scientists from the tedious and inefficient tasks of information gathering and cross-validation,epiXora™ reconstructs the digital foundation of biopharmaceutical research through massive domain-specific data, a logic framework aligned with real-world decision chains, and rigorous verification mechanisms:

• At the data level,epiXora™ covers 56 million literature metadata entries and 9.4 million full-text articles, and integrates multi-source professional databases including clinical trials, regulatory documents, market sales data, and 160 million patent records—comprehensively breaking down barriers posed by massive, heterogeneous data and providing a solid evidentiary foundation for subsequent multidimensional scientific analysis.

• At the logic level,epiXora™ embeds industry-standard research frameworks and decision logic, automatically structuring research content around scientific mechanisms, disease relevance, clinical progress, competitive landscape, patent risks, and commercial potential. Users can interact via "Quick Mode” to receive a structured briefing within 5 minutes and form an initial judgment; or enter “Deep Mode” to generate a comprehensive, multi-dimensional report in approximately 30 minutes, significantly reducing post-processing costs and aligning the research path more closely with real-world decision-making.

• On the credibility front,, epiXora™ is built on a foundation of trustworthiness. Key assertions supportsentence-level source tracing,enabling one-click retrieval of original source passages. Meanwhile, to address the common 'hallucination' problem of large models in professional applications, epiXora™’shallucination detection feature，flags high-risk content with inconsistent evidence or missing sources, helping researchers quickly verify findings. Research materials and conclusions can be stored in a structured format withinKnowledge Map, enabling reuse in subsequent projects and transforming one-off research outputs into continuously accumulated knowledge assets.

epiXora™ has undergone comprehensive iteration based on user feedback from its predecessor, X-Buddy,and is now open for trial access by professionals in the life sciences and pharmaceutical R&D fields, offering a scalable, intelligent solution for traditionally manual-intensive expert research workflows, while providing stronger front-end analytical support for XtalPi’s integrated 'AI + Robotics + Multi-Agent' R&D framework. Within this framework, specialized scientific agents—equipped with rigorous logic and autonomous reasoning capabilities—are empowering critical stages ranging from target analysis and molecule generation to automated synthesis experiments, driving AI’s evolution from an 'efficiency-assisting tool' toward a 'professional research partner' capable of multi-step collaboration.

From answering a single expert scientific query to building a continuously evolving, highly trustworthy intelligent network, epiXora™ is redefining the efficiency and certainty of professional research through its life science–dedicated AI agent capabilities. In the future, XtalPi’s purpose-built Multi-Agent system will continue advancing deep AI integration across core scientific research scenarios, establishing intelligence-driven scientific discovery as the new baseline for industry advancement.

42K Views