Integrate, manage, secure, and analyze all of your enterprise data.

The Platform

On the back-end, the Palantir Gotham platform comprises a suite of capabilities for integrating many different data sources for secure, collaborative analysis. The platform serves as an enterprise knowledgebase, containing the full record of an organization's collective analysis.

Flexible Modeling

Instead of rigid rows and columns, Palantir Gotham models data as a flexibly specified graph of objects and relationships—real-world entities like "people" and "organizations," their relevant features, and the connections between them.

The Palantir Gotham data model, called the "Dynamic Ontology" because it can be defined and re-defined on the fly, makes it possible to integrate many different kinds of data from many different sources into a coherent whole that reflects how people naturally conceive of information.

Privacy and Security Controls

Privacy-protective capabilities are built into the platform's very architecture, which is designed to support precision data handling, multi-level security, and complete auditability. First, every property of every object integrated into the platform is tethered to its original data source, where access restrictions can be applied on a per-property basis (sometimes referred to as "sub-cell level security").

Then, users can be assigned a variety of access permissions that govern their ability to interact with this data. Finally, all user and administrator interaction with data is recorded in a tamper-proof audit log.


The Revisioning Database and Palantir Gotham's Nexus Peering technology enables multiple users, within and across organizations, to seamlessly, securely, collaboratively analyze the same data. The Palantir Gotham platform supports collaboration across organizational, functional, and geographic boundaries; across security models and data models; and across low-bandwidth, high-latency networks, including satellite connections all while preserving data security and integrity.

Extensibility, Customizability, and APIs

The Palantir Gotham platform is designed to be extensible at every layer of the stack. From low-level data integration, import pipeline customizations, to building custom user interface, it has been designed as a fundamentally open platform. Data that has been integrated via the Dynamic Ontology can be accessed as Palantir Objects via our Java API. Data can also be exported wholesale for use in other frameworks and tools.

Knowledge Management

All integrated data is stored in Palantir Gotham's Revisioning Database. The RevDB tracks every change made to an object, whether these changes originate in the source data or are made by users. Conceptually similar to Git and other distributed version control systems, the RevDB allows analysts to work in individual sandboxes, which are tracked using a branching history of revisions, until they are ready to publish their findings to the enterprise.

Users can explore divergent lines of reasoning, record each step along the way, and jump back to earlier points in their investigations. Analysts can share insights without losing their own work. The result is a version-controlled knowledgebase that represents the accumulated insights of an organization's users, turning their analyses into data that can be further leveraged by the enterprise.

Algorithmic Processing

Built-in algorithmic capabilities augment the human user's ability to make sense of large-scale data by automatically identifying interesting clusters of data to queue up for analysts to review. Palantir Phoenix serves as the place where massive datasets can be compiled and analyzed. It also delivers a powerful and flexible framework with which to implement automation. Non-technical analysts can use our seed-generation framework to graphically author new jobs without writing a line of code.


The platform handles petabyte-scale data through a combination of scalable architecture and federated data storage. Our Phoenix server acts as a data warehousing system to hold large structured datasets like log files, network traffic flows, and transactional data. Our Raptor federated search server holds large collections of unstructured text data like documents, emails, and cables.

Both make their data immediately available to users via searching and other enhanced querying helpers. When users request data from Phoenix or Raptor, it is integrated into the RevDB for analysis.