Tools, Models & Compute Modules
Tools we install, models we've trained, and governed compute that runs inside your platform. Built close to operators, validated in production, and designed to snap into Palantir Foundry or run independently in your own cloud. Filter by type, use case, or industry.
Tool
Type
Input / Output
Where it fits
Industries
External React/Next.js app layer for Palantir Agent Studio agents, built with Ontology SDK and Platform SDK.
User Query + App Context→Grounded Answer + Agent Actions
Complete setup for YOLO object detection model inference directly inside Foundry.
Foundry Media Sets→Bounding boxes + confidence
Manages context-specific model variations without creating entirely new tools.
Base Model→Context-Specific Weights
Paddle-driven detection pipeline to find regions of interest and output clean crops for downstream tools.
Image / PDF→Bounding boxes + cropped regions
Extracts rows, columns, and cell structure so messy tables become parseable.
Table Image / PDF→Table Structure + Coordinates
Specialized OCR model for extracting and structuring data from handwritten tables in legacy documents.
Image / PDF→Text + confidence
Compares two signatures to validate authenticity with a similarity score.
Signature A + B→Similarity score
Lightweight human-in-the-loop review and correction interface with audit trail.
AI output→Human-verified data
Row/column matching across datasets with exception handling.
Dataset A + B→Matches + confidence
Secure, controlled write-back to systems of record.
Validated data→ERP update + audit
Uses language models to filter low-value retrieval results and improve grounded generation quality.
Vector Results→Relevant Context
Evaluates AI output and routes low-quality results back for regeneration.
AI Draft→Improvement Feedback
Interactive graph view of linked objects, relationships, and operational context.
Ontology Objects + Links→Graph View
External React/Next.js app layer for Palantir Agent Studio agents, built with Ontology SDK and Platform SDK.
Input / Output
User Query + App Context→Grounded Answer + Agent Actions
Complete setup for YOLO object detection model inference directly inside Foundry.
Input / Output
Foundry Media Sets→Bounding boxes + confidence
Manages context-specific model variations without creating entirely new tools.
Input / Output
Base Model→Context-Specific Weights
Paddle-driven detection pipeline to find regions of interest and output clean crops for downstream tools.
Input / Output
Image / PDF→Bounding boxes + cropped regions
Extracts rows, columns, and cell structure so messy tables become parseable.
Input / Output
Table Image / PDF→Table Structure + Coordinates
Specialized OCR model for extracting and structuring data from handwritten tables in legacy documents.
Input / Output
Image / PDF→Text + confidence
Compares two signatures to validate authenticity with a similarity score.
Input / Output
Signature A + B→Similarity score
Lightweight human-in-the-loop review and correction interface with audit trail.
Input / Output
AI output→Human-verified data
Row/column matching across datasets with exception handling.
Input / Output
Dataset A + B→Matches + confidence
Secure, controlled write-back to systems of record.
Input / Output
Validated data→ERP update + audit
Uses language models to filter low-value retrieval results and improve grounded generation quality.
Input / Output
Vector Results→Relevant Context
Evaluates AI output and routes low-quality results back for regeneration.
Input / Output
AI Draft→Improvement Feedback
Interactive graph view of linked objects, relationships, and operational context.
Input / Output
Ontology Objects + Links→Graph View
13 of 13 tools