AIPOCH logo

AIPOCH

Open-source AI research workbench with 550+ medical research agent skills for evidence synthesis, protocol design, data analysis, and academic writing.

11-50 employees
Visit website
AI-generated from publicly available materials.Is this your company? Claim your profile

Overview

AIPOCH is a Singapore-based company (legal entity AIPOCH Pte. Ltd.) that develops open-source AI tools for scientific and medical research. Led by CEO Huimei Wang, the company describes its offering as an "open-source harness for scientific research," combining a model-agnostic research workbench with a library of over 550 medical research agent skills spanning evidence synthesis, protocol design, data analysis, and academic writing.

AIPOCH's tools are designed to serve researchers and clinicians working across scientific and medical domains, with a focus on open, local-first deployment and interoperability across AI models.

Core Products

  • Open-Science: The company's flagship product, Open-Science is an open-source, local-first AI research workbench. It serves as the central platform through which AIPOCH's agent skills and research capabilities are accessed.
  • MedFlow: An announced product focused on supporting clinical research workflows, extending AIPOCH's tooling into more structured clinical research processes.
  • Evova: An announced evidence evaluation platform, intended to support the assessment and synthesis of scientific evidence.

Agent Skills Library

  • AIPOCH maintains a library of more than 550 medical research agent skills available through its workbench.
  • Skills cover a range of research activities including evidence synthesis, protocol design, data analysis, and academic writing.
  • The workbench is described as model-agnostic, meaning the agent skills are not tied to a specific underlying AI model.

Quality Assurance and Research

  • To assess the quality of its agent skills prior to deployment, AIPOCH developed MedSkillAudit, a pre-deployment audit framework.
  • MedSkillAudit was created in collaboration with the Department of Pathology at Zhongshan Hospital, Fudan University.
  • The framework was published as an arXiv preprint in April 2026, indicating an academic and peer-facing approach to validating its tooling.

AIPOCH's work reflects an approach to AI-assisted research that emphasises openness, local-first deployment, and structured quality evaluation, with academic collaborations informing the development and validation of its core capabilities.