HistAI

HistAI develops digital pathology software and data infrastructure for pathologists, researchers, and pharmaceutical and biotech organizations. The company operates the CellDX platform, which combines a large curated whole slide image (WSI) dataset, cloud-based slide viewing and annotation tools, and proprietary AI models for computational pathology. HistAI is also a contributor to open-source pathology research, publishing datasets and foundation models on Hugging Face. The platform is used by over 3,000 researchers at leading organizations.

HistAI positions itself as a provider of the world's largest WSI data hub, offering commercially licensed pathology images with transparent, volume-tiered pricing and near-instant cohort access. Its infrastructure runs on Microsoft Azure, and its open-source contributions include over 112,000 published WSIs, 5.3 million manually labelled images, and two foundation models released under the Apache 2.0 license.

CellDX Platform Modules

• Data Hub: A curated repository of whole slide images available for purchase with perpetual commercial licenses. As of early 2026, the hub contains over 212,000 slides, including 48,333 IHC slides. Images span a wide range of tissues, stains, clones, and diseases — including H&E and IHC stains across organs such as lung, breast, cervix, kidney, liver, skin, and brain. Cohorts can be filtered, statistically previewed before purchase, and downloaded within minutes of payment.
• Slide Viewer: A cloud-based viewer for seamless exploration of whole slide images, supporting collaborative access by multiple users simultaneously.
• Annotation Tools: A set of manual annotation tools for marking and labelling slides, with support for simultaneous multi-user annotation workflows.
• Proprietary AI Models: A growing library of specialized pathology AI models for tasks such as cancer detection. An example output shown on the platform includes ductal cell carcinoma detection with 94.2% confidence.
• Data Hub Skills: Launched in February 2026, this toolset provides autonomous programmatic access to over 160,000 WSIs and is compatible with AI coding agents including Claude Code, Gemini, and CODEX.

Target Users and Use Cases

• Pathologists: Access to AI-assisted analysis workflows and intuitive slide management tools intended to reduce manual review time.
• Researchers: Access to large, diverse WSI datasets to support computational pathology studies and accelerate discovery workflows.
• Pharma and Biotech: AI models trained on diverse patient cohorts to support biomarker development and multi-cohort studies. The platform supports analysis across multiple patient groups simultaneously.

Open-Source Contributions and Publications

• HistAI has published over 112,000 WSIs on Hugging Face, described as the largest publicly available WSI dataset.
• Two foundation models for pathology are available under the Apache 2.0 license.
• Over 5.3 million manually labelled images have been made publicly available.
• HISTAI dataset paper (arXiv, May 2025): Describes an open-access collection of over 100,000 multimodal whole slide images with rich metadata, authored by Dmitry Nechaev, Alexey Pchelnikov, and Ekaterina Ivanova.
• SPIDER dataset paper (arXiv, April 2025): Presents the largest open-access patch-level dataset spanning multiple organs, with expert annotations and baseline models.
• Hibou foundation models paper (arXiv, August 2024): Describes a family of vision transformers pretrained on over 1 million whole slide images.

Partnerships and Infrastructure

• In October 2025, HistAI announced a partnership with Protege to deliver one of the largest whole-slide pathology datasets to AI developers.
• Cloud storage and infrastructure are hosted on Microsoft Azure.
• WSI pricing is tiered by volume and stain type, with H&E images ranging from $2.00 to $5.00 per slide and IHC images from $23.00 to $40.00 per slide depending on quantity purchased.

HistAI serves the full spectrum of pathology stakeholders through a combination of commercially licensed data, collaborative cloud tools, and open-source model development, with an active publication record supporting its research-oriented positioning.