• Documentation
    • Search documentation
    karat

    +

    K

    API Reference ↗
    Platform overviewAIP capabilities

    AIP capabilities

    Applications across the Palantir platform are equipped with AIP-powered capabilities. This page describes a number of these capabilities.

    AIP's AI functionality can be divided into three categories:

    • AIP Assist: An LLM-powered support tool designed to help users navigate, understand, and generate value with the Palantir platform. Users can ask AIP Assist questions in natural language and receive real-time help with their queries.
    • AIP assistant features in platform applications: Native LLM-backed features designed to help end users perform regular workflows in the Palantir platform. These are highly-specific features that leverage knowledge of the platform to accelerate a user's day-to-day operations.
    • AIP capabilities for custom workflows: A set of capabilities that allow developers to build their own LLM-backed workflows or applications. These are open-ended functionalities built for developers or data scientists.

    Learn how to enable AIP features in the platform.

    Understand the difference between AIP core features and AIP capabilities for custom workflows.

    AIP Assist sidebar

    The AIP Assist sidebar is context aware and can provide users support using natural language prompts. You can open AIP Assist from your workspace navigation bar or access it with a keyboard shortcut (Cmd + Shift + U (macOS) or Ctrl + Shift + U (Windows)). AIP Assist will appear in a panel as shown in the screenshot below.

    AIP Sidebar overview screenshot.

    Learn more about the AIP Assist sidebar.

    AIP features in platform applications

    AIP features have been embedded into core applications to help users accelerate their workflows and unlock more value in the platform. The following are a selected sample of AIP features and is a non-exhaustive list. The latest updates on AIP can be found in the Announcements section of the documentation.

    Pipeline Builder

    Use AIP in Pipeline Builder to help you better understand, build, and manage your pipeline. Pipeline Builder has a core set of Assist features and additional AIP capabilities for custom workflows.

    A sample of core Assist features in Pipeline Builder include:

    Explain: Learn more about the steps in your pipeline development and suggest relevant names and descriptions.

    AIP Explain feature in Pipeline Builder graph, explaining the relationship of 13 nodes.

    Regex Helper: Generate a RegEx customized for your needs, suitable for all skill levels.

    Using the Regex Helper feature in Pipeline Builder to generate a regular expression to search for email domains.

    Transform Assist: Create and edit regular expressions and easily cast strings to specific timestamp formats.

    Additionally, with permission, you can use AIP capabilities for custom workflows in Pipeline Builder, such as the following:

    The Use LLM node, offering a convenient method for executing large language models (LLMs) on your data at scale. Five pre-engineered templates are provided for a beginner-friendly start to using LLMs that leverages the expertise of experienced prompt engineers.

    The "Create a prompt" screen for the Use LLM node in Pipeline Builder.

    You can also run trials over a few rows of your input dataset to iterate on your prompt before running your model on an entire dataset.

    Trial runs for the Use LLM Node in Pipeline Builder.

    Notepad

    AIP also brings LLM-powered functionality to Notepad, where you can use AIP to automatically spellcheck, shorten, modify, or translate your text without affecting the document's existing formatting.

    AIP dropdown menu in Notepad showing available features.

    Scheduler

    You can use AIP in the Scheduler application to generate a schedule configuration when creating a dataset build schedule with a specific time trigger. Input a schedule trigger prompt within the New schedule view side panel to generate the proper cron format for complex triggers.

    AIP Threads [Beta]

    AIP Threads enables users to accomplish a variety of tasks and ad-hoc analyses with the power of LLMs. No setup or technical expertise is required to interact with documents (for example, PDFs) and AIP Agents (interactive assistants equipped with enterprise-specific information and tools). To begin, you need only drag and drop your documents into the interface, pick from previously uploaded documents that you have access to, or select an AIP Agent that you and your organization have created.

    The main AIP Threads interface with a number of conversations with documents and AIP Agents. The selected conversation asks a technical question and the answer includes a citation to the original document.

    AIP capabilities for custom workflows

    AIP capabilities for custom workflows allow developers and builders to build their own LLM-backed workflows or applications in the Palantir platform. These include but are not limited to AIP Logic, Pipeline Builder "Use LLM" nodes and text-to-embeddings, AIP Automate, AIP Agent Studio, and AIP Workshop widgets. These capabilities natively support large language model options.

    Palantir-provided LLMs are also available in core Foundry features such as Functions, Transforms, and Jupyter® notebooks via Code Workspaces.

    In addition, existing Palantir capabilities for model integration allow users to connect custom large language models and independently build use cases from zero.

    Platform administrators may govern usage of these capabilities via Control Panel under AIP settings.

    Supported LLMs

    The Palantir platform provides support for a number of LLMs (large language models) and text embedding models.

    Palantir-provided large language models (LLMs)

    We have made a set of LLMs accessible for use with AIP capabilities for custom workflows, with LLM selection and availability differing across enrollments. Details about the available models can be found below:

    To understand how LLMs securely process user prompts, review FAQs: Security and Privacy of Palantir's AIP leveraging third-party hosted LLMs ↗ by selecting Palantir AIP FAQs.

    Text embedding models

    We have also made a set of text embedding models available in the same way.

    Learn how to configure which LLMs are available on your enrollment.

    Note: AIP feature availability is subject to change and may differ between customers.

    Jupyter®, JupyterLab®, and the Jupyter® logos are trademarks or registered trademarks of NumFOCUS.

    The "OpenAI" name and the “GPT” brands are property of OpenAI.

    All third-party trademarks (including logos and icons) referenced remain the property of their respective owners. No affiliation or endorsement is implied.