Parallel Domain Launches PD Forma Bringing 4D Reconstruction, Quality Assessment and Scenario Extraction into One Pipeline for Tremendous Speed, Scale and Fidelity

PD Forma enables program scale reconstruction of simulation-ready environments in hours as autonomous systems use cases expand

SAN FRANCISCO, Nov. 4, 2026 /PRNewswire/ -- Parallel Domain, a leader in sensor simulation and neural reconstruction, today announced the launch of PD Forma, software that transforms raw, real-world capture data into validated, simulation-ready environments in hours. Powered by the firm's 4D Geometric Foundation Model, PD Forma is designed to help autonomy teams move faster, detect and solve failures early and simulate the complexity of the real world with confidence.

Building on its PD Replica and PD Sim products, PD Forma is the end-to-end platform that brings reconstruction, quality assessment and scenario extraction into one pipeline. With PD Forma, autonomy teams scale their simulation workflows faster and without the engineering bandwidth traditionally required.

PD Forma gives teams the full workflow to build replicas on their own through a web app, CLI or API. Teams validate and curate raw logs before upload, then process them in batches using presets for driving, drone and indoor scenes. When a run fails, stage logs show what broke, and teams restart from that stage instead of starting over.

Every replica comes with a quality report that compares the rendered output to the original capture, which can be reviewed before publishing the replica to PD Sim for camera, lidar and radar simulation.

Solving the testing roadblock as AI-enabled systems expand

With autonomy expanding rapidly into more environments, developers are pressed with building systems that perform reliably beyond the conditions they have already encountered. Real-world data provides a foundation but falls short of exposing an AI model to the range of elements it may encounter, from weather to lighting to road conditions and more. Teams must be able to reproduce real-world scenarios, generate new variations and use simulation throughout development, training and validation.

PD Forma streamlines the process of converting raw capture data into high-quality, simulation-ready environments and does so with 4D capabilities, allowing reconstruction of not only the static environment but also dynamic agents and behaviors within it.

"With PD Forma, teams can move from real-world logs to validated, simulation-ready environments in just hours without sacrificing fidelity," said Zack Novak, Parallel Domain CEO. "We're helping customers accelerate their time to market by making program scale simulation more accessible and efficient. And we've seen demand explode; customer usage has increased 19X over the past six months."

Diagnosing failures without starting over

When a reconstruction encounters an issue, teams can use staged logs, quality reports and the ability to restart from a specific stage rather than starting over from scratch. The output is an inspectable, simulation-ready reconstruction faster. Each replica comes with a quality report so teams can review reconstruction quality before approving and publishing it to PD Sim for simulation workflows.

PD Forma also allows for the reuse of the same reconstructed scene across different times of day and weather conditions such as dawn or dusk, rain or snow. By preserving the scene's geometry while varying the environmental conditions, teams can evaluate how their systems respond to changing conditions.

"By reconstructing the scenario with variations, PD Forma is designed to help users generate simulation experiences that pressure test how things may change as the environment does," said Novak. "That more effectively supports AI training, enables the investigation of system behavior and validates performance across conditions that would be difficult to capture consistently in the real world."

One platform for multiple simulation needs and faster deployment

PD Forma is built to support a broad range of simulation needs beyond the recreation of driving scenes, including:

• Drone delivery: Turn flight logs into replicas of the neighborhoods a drone serves, then test every approach and drop-off at dawn, at dusk, in rain or in snow.
  
  
• Autonomous aircraft: Reconstruct airfields and approach paths so teams can test perception on landing without flying the same approach hundreds of times.
  
  
• Robotics: Recreate the warehouses and facilities robots work in so teams can test around people and equipment before they deploy.

Teams can also incorporate multiple capture logs into larger replicas, work with mixed-camera inputs and dynamically reconstruct a scenario, enabling teams to replay a captured scenario or vary it in simulation.

PD Forma is the newest part of Parallel Domain's broader AI-driven simulation platform, connecting real-world capture data with high-fidelity environments and dynamic, 4D scenario generation. The platform enables teams to incorporate simulation across development, validation and training, helping autonomous systems learn and improve over time.

To learn more, visit www.paralleldomain.com.

About Parallel Domain

Parallel Domain is a global leader in scene reconstruction and realistic sensor simulation to develop, test, and validate autonomous systems. Its proprietary scene reconstruction from drive log and simulation technologies deliver photorealistic, labeled sensor data on demand, accelerating safe autonomy with realism that scales. Learn more at www.paralleldomain.com.

Media contact:

Megan Lampros
One Twelve Communications on behalf of Parallel Domain
megan@onetwelvecomms.com

SOURCE Parallel Domain