Skip to main content

PARVIS

Validate Vehicle APIs on a virtual vehicle before the real one exists

PARVIS is an AI-powered Vehicle API simulation platform for SDV application development. Apply SDV Vehicle APIs to a web-based 3D virtual vehicle, and build and test third-party apps before they go into a real car.

Contact us
The real PARVIS screen. The app keeps calling Vehicle APIs while the car drives, and the right panel shows which candidate path the AI chose.

From Vehicle API to third-party app, in one flow

Early in development there is no physical vehicle. PARVIS connects the whole SDV development flow around a virtual vehicle.

  1. 01

    Vehicle API

    Calls, responses and events run exactly as the standard SDV Vehicle API defines them.

  2. 02

    3D virtual vehicle

    Doors, windows, trunk, lamps and driving move on the 3D screen as the APIs are called.

  3. 03

    Adaptive / Classic AUTOSAR vECU

    A setup that connects virtual ECUs based on Adaptive AUTOSAR and Classic AUTOSAR, configured for each project.

  4. 04

    Third-party apps

    Build and test apps that use Vehicle APIs on the virtual vehicle.

The SDV development flow across PopcornSAR products

Within the PopcornSAR product line, PARVIS covers the Vehicle API and the virtual vehicle. From requirements and design to virtual ECUs, Vehicle APIs and third-party app testing, the flow runs on one company's tools.

  1. 01Requirements · verification work products
  2. 02AUTOSAR design
  3. 03Virtual ECU (vECU)
  4. 04Vehicle API · 3D virtual vehicle
    PARVISThis page

    Doors, windows, trunk, lamps and driving move on the 3D screen as the APIs are called.

  5. 05Third-party app testing
    PARVISThis page

    Build and test apps that use Vehicle APIs on the virtual vehicle.

Can a standardized Vehicle API really be applied to any vehicle design?

We tested this question with a Tesla Cybercab in PARVIS. We applied the Vehicle APIs defined by two SDV standardization activities to the same vehicle and checked which APIs can be implemented.

K-SDV (Korea)
40 / 57
40 of 57 convenience API attributes applicable
OSDVI (Japan)
15 / 25
15 of 25 API objects applicable

The initial validation was completed within one day.

Even when APIs are standardized, what actually applies depends on the vehicle's physical architecture, actuators, sensors and functions. A car with cameras instead of side mirrors, or no wipers, cannot implement those APIs.

The real PARVIS screen. Devices the real car does not have are not modeled, and their APIs answer as not supported.
Vehicle API validation should begin together with vehicle architecture and design — not after the vehicle has already been designed.

What you can simulate with PARVIS

Reproduce a wide range of vehicle conditions in a virtual 3D environment.

  • Vehicle control and convenience functions

    Control doors, windows, trunk, mirrors, wipers and lamps through Vehicle APIs, and check how responses differ while driving and while stopped.

  • Autonomous driving scenarios

    The AI drives through surveyed city maps, traffic signals, other vehicles and pedestrians.

  • Autonomous parking scenarios

    Drive into a lot and reverse into a bay automatically; completion is judged by position and heading error.

  • Virtual ECU integration

    A setup that connects vECUs based on Adaptive and Classic AUTOSAR, configured for each project.

  • Third-party app scenarios

    Reproduce an app service that uses Vehicle APIs end to end, from hailing a robotaxi to boarding, alighting and parking.

Autonomous parking: reversing into the bay, seen from above.
Robotaxi scenario: hail from the passenger app, board, then ride to the destination lot.

Lightweight web 3D, so it is fast

No high-performance GPU needed
Unlike simulators based on Unreal Engine or Unity, PARVIS uses a JavaScript-based web 3D architecture, so Vehicle APIs can be validated even on relatively low-spec PCs.
Change it and test again right away
Vehicle models, APIs, behaviors and scenarios can be changed and tested quickly without rebuilding a heavy simulation environment.
Natural-language setup with an AI harness
Engineers set up and explore SDV and AI driving scenarios in natural language. AI records the results automatically during the simulation.

Standard Vehicle APIs, and each automaker's own

Developed for an actual customer project in Japan, PARVIS implements both the Korean and the Japanese Vehicle API standards. Both are drafts under development, and PARVIS is updated as new revisions arrive. Beyond the standards, Vehicle APIs defined by individual automakers are added and supported the same way.

K-SDVKorea

Korea is standardizing SDV Vehicle APIs. 2026 focuses on convenience functions, followed by energy, ADAS and autonomous driving APIs in 2027.

Convenience: 10 modules · 57 attributes

OSDVIJapan

The SDV API of the Open SDV Initiative. Body and ADAS/AD objects run exactly as the source defines them.

25 objects · 194 service calls

OEM Vehicle APIPer automaker

We take the automaker's own Vehicle API specification, run it exactly as written just like a standard, and apply it to a virtual vehicle that matches that automaker's vehicle configuration.

Configured per project

Specifications

Runs in
A web browser (WebGL 3D); no game engine to install
Standards
K-SDV (Korea), OSDVI (Japan) — drafts of each standard under development. Automaker-specific Vehicle APIs (configured per project)
Scenarios
Vehicle control and convenience functions, autonomous driving, autonomous parking, third-party apps such as robotaxi
Integration
Adaptive / Classic AUTOSAR vECU (configured per project)
Screen languages
Multilingual, including Korean, Japanese and English

What vehicle or mobility service would you want to simulate first?

We will review the vehicle configuration and Vehicle API applicability with you.

Contact us