Anyverse ADAS & AD
Anyverse ADAS & AD is a web application that generates high-fidelity synthetic datasets for training, testing and validating advanced driver-assistance and autonomous driving systems, simulating cameras, LiDAR, radar and thermal sensors with physically accurate light transport.
What is Anyverse ADAS & AD?
Anyverse ADAS & AD is the synthetic data application that Anyverse S.L. sells to teams building advanced driver-assistance and autonomous driving systems. It is a web application: engineers configure scenarios, sensors, variability and outputs in a browser interface, without writing code, and it shares its interface and workflow with the company's separate in-cabin product.
What sets it apart is the engine underneath. Rather than a game engine, Anyverse runs a proprietary spectral renderer that simulates light transport with physical accuracy, which the vendor presents as the reason its data transfers to the real world instead of widening the sim-to-real gap. Sensors are modelled with their own physics: RGB cameras, near-infrared, infrared and thermal imaging, LiDAR and radar all respond to the same scene as their real counterparts would to light, materials and weather.
The published use cases stay close to what validation teams actually test. Lane keeping and lane departure warning, pedestrian and cyclist detection, traffic sign and light recognition, vehicle detection and classification, adaptive cruise control, automatic emergency braking, sensor-fusion scenarios for Level 2 and Level 3 autonomy, plus intersections, overtaking and merging. Scenes draw on a built-in asset library of vehicles, pedestrians and cyclists with varied demographics, buildings, street furniture, vegetation, construction barriers, debris and animals, placed in urban, suburban, rural, highway, tunnel, roundabout, car park or off-road settings, including geo-located or invented layouts.
The value for a machine learning pipeline lies in the outputs. Every render can carry up to fourteen Arbitrary Output Variables, among them semantic and instance segmentation, material maps, depth, 3D position, surface normals, roughness, albedo, radiance, spectral radiance, raw sensor images and 2D and 3D motion vectors, delivered in the formats and ontology the customer's stack expects. Rendering is distributed by a cloud engine and datasets are stored in the cloud. Anyverse also ships scenario libraries aligned with Euro NCAP protocols, which are the benchmarks its customers must satisfy rather than certifications the vendor holds.
What it does
- Generate high-fidelity synthetic driving datasets with pixel-accurate ground truth
- Simulate RGB, near-infrared, infrared, thermal, LiDAR and radar sensors with their physical properties
- Recreate corner cases such as fog, glare, hail, smoke, night transitions and wet reflective roads
- Export up to fourteen annotation layers, from semantic and instance segmentation to spectral radiance
- Build scenario sets aligned with Euro NCAP and UNECE test protocols
- Model sensor failures, occlusions and cross-talk without risking a vehicle or a driver
- Scale rendering through a cloud engine from hundreds to millions of annotated samples
When to use Anyverse ADAS & AD / When not to
A quick filter to help you decide if Anyverse ADAS & AD is the right fit.
When to use Anyverse ADAS & AD
- Perception engineers at car makers and Tier 1 suppliers who must validate ADAS functions before physical prototypes exist
- Validation teams preparing campaigns aligned with Euro NCAP and UNECE test protocols
- Sensor-fusion specialists who need camera, LiDAR, radar and thermal data captured of the very same scene
- Machine learning teams starved of rare or dangerous corner cases that cannot safely be filmed on real roads
- Computer vision research groups needing pixel-perfect ground truth that manual annotation cannot deliver
When not to use Anyverse ADAS & AD
- Anyone hoping to sign up online and see a price: every route runs through a sales form and a negotiated agreement
- Teams that need a documented public API or SDK to drive dataset generation from their own code
- Small budgets looking for a permanent free tier, since free access is limited to a time- or capacity-bound evaluation
- Creative users after a general-purpose image generator, as this engine is built for machine vision, not illustration
- Buyers focused on in-cabin driver and occupant monitoring, which Anyverse sells as a separate product with its own terms
How to use Anyverse ADAS & AD
A typical end-to-end flow, from setup to results.
- Identify which sensors, functions and edge cases your validation plan actually needs, since nothing is self-service
- Request a demo through the ADAS demo form, or use the general contact form to ask about a subscription or a trial
- Discuss scope with the Anyverse team, who tailor the setup to your development stage
- Agree commercial terms, which set your number of licensed users and your rendering capacity
- Receive unique credentials and sign in to the web application
- Define your taxonomy, ontology and labelling specification so outputs land in your pipeline without reformatting
- Configure scenarios, environments, sensor placements, lighting and weather in the interface, without coding
- Select which of the fourteen annotation layers you want exported
- Launch procedural generation and let the cloud engine distribute the rendering workload
- Collect datasets from the cloud, then iterate on new corner cases as your models expose weaknesses
Pros & Cons
Pros
- Physically accurate spectral rendering rather than a repurposed game engine, aimed squarely at reducing the sim-to-real gap
- Unusually wide sensor coverage, with radar and thermal alongside the more common camera and LiDAR
- Fourteen layers of ground truth delivered with every frame, which no manual annotation effort could match
- Contract terms on generated data are strongly in the customer's favour: perpetual, worldwide and exclusive rights
- Version 3 of the licence adds explicit protection for confidential customer inputs and derived assets
- Anyverse holds TISAX certification, the automotive industry's own information security assessment, announced in March 2026
- Hosting is documented and four named subprocessors are published, which is more transparency than most vendors offer
Cons
- No public pricing whatsoever: no pricing page exists, and none of roughly thirty-five probed paths or the five sitemaps reveals a figure
- No permanent free plan, only an evaluation licence capped by time or rendering capacity and barred from commercial use
- No documented product API or SDK, and the licence expressly forbids automated access to the products
- No Data Processing Agreement is published or offered, despite the vendor handling customer material
- Nothing is published about whether customer data is used to train models, and no opt-out mechanism is documented
- The product terms are a downloadable PDF rather than a readable web page, and the page linking it is absent from the site navigation
- Web only, with no mobile application and no documented on-premise deployment for this offering
Pricing & Plans
There is no permanent free plan and no public price point of any kind. Anyverse publishes no pricing page: the five sitemaps list none, and around thirty-five candidate paths all return genuine 404 responses. The company sells through a form, a demonstration and a negotiated agreement. The licence agreement does describe the commercial mechanism: charges take the form of a single payment, a subscription fee and/or rendering capacity, with the licence granted per user and coupled to a defined rendering capacity that can also be bought as prepaid credit. Maintenance and email support are included with a subscription, and with prepaid capacity for twelve months after the last purchase. A free-of-charge evaluation licence exists, limited in duration or capacity and restricted to non-commercial assessment. Because no amount is published, no starting price, currency or billing unit can honestly be stated.
Data, GDPR & hosting
A consolidated view of how Anyverse ADAS & AD handles your data.
GDPR overview
The site makes concrete, checkable GDPR statements rather than a vague compliance badge. The controller is named as Anyverse, S.L., VAT ESB81938813, Paseo de la Castellana 200, 28046 Madrid, reachable at gdpr@anyverse.ai. Each processing purpose is tabled against its legal basis, whether regulatory obligation, contract performance, consent or legitimate interest. The full set of data subject rights is listed, with access, rectification, erasure, restriction, portability, objection and withdrawal of consent, exercised in writing with proof of identity or by email. Transfers outside the European Union rely on the Commission's Standard Contractual Clauses or on adequacy decisions, and four recipients are named. The supervisory authority is identified as the Agencia Espanola de Proteccion de Datos. Parental consent is required below the age of fourteen. No data protection officer and no Article 27 representative are designated, which is consistent with a controller established inside the Union.
Who owns the data?
The End User License Agreement draws a sharp line. Synthetic data produced by the service is licensed to the customer royalty-free, perpetual, worldwide, unlimited, for any purpose and exclusively, though the rights cannot be transferred or sublicensed. Anything the customer supplies, such as non-public CAD files or proprietary specifications, stays the customer's "Client Materials", and deliverables built from them become "Custom Client Assets" that Anyverse undertakes never to disclose, license or commercialise to third parties, beyond staff, contractors and hosting providers bound by confidentiality. The software, the render engine and the generic scene libraries, called "Anyverse Standard Materials", remain the sole property of Anyverse S.L. and stay reusable for other customers.
Reuse rights
Customers may use the synthetic datasets they generate without asking further permission: the licence is perpetual, worldwide and unrestricted as to purpose, which covers internal training, testing, validation and shipping the resulting models. Three limits apply. The data cannot be resold, redistributed or made downloadable as standalone files apart from the project it serves, it cannot be used unlawfully or in breach of industry codes, and it cannot become part of a trademark, trade name or logo. Merely configuring standard scenes, sensor positions or lighting does not turn Anyverse's own reusable assets into customer property. On the vendor side, the agreement commits Anyverse to process customer-specific assets solely to deliver the service, and neither the licence nor the privacy policy says anything about using customer data to train AI models.
Data retention & training
Hosting summary
Anyverse states that it runs its service infrastructure on third-party cloud providers, naming Google Cloud Platform, and that it applies encryption at rest and in transit, regular backups and redundancy, while acknowledging that availability depends on those providers. The privacy policy publishes a transfer register naming four recipients: Google in Ireland for credential storage, MongoDB Atlas in Ireland for platform credentials, Atlassian in Ireland for internal task management, and AFI Intl Inc in the Netherlands for backup, each covered by EU Standard Contractual Clauses. The jurisdiction is Spanish, with Spanish and European data protection law applying and the Agencia Espanola de Proteccion de Datos as supervisory authority. One inconsistency deserves attention: the licence agreement allows personal data to be stored and processed in the European Union, the United States or any other country where the vendor has facilities or engages vendors, which is broader than the transfer register suggests. A network lookup places the website itself on Google infrastructure in Madrid, though the site does not state this.
Things to keep in mind
Risks and trade-offs to weigh before adopting Anyverse ADAS & AD.
- Regulatory name-dropping is easy to misread: Euro NCAP, UNECE, ISO 26262 and SOTIF are the benchmarks the data helps you satisfy, not certifications Anyverse holds. Only TISAX is the vendor's own.
- The terms linked in the site footer cover the separate in-cabin application only. The agreement governing this offering is a PDF reached through a page that is absent from the navigation menu.
- The eighteen-year age requirement belongs to the in-cabin product's terms. No age condition is published for the ADAS and AD offering.
- The documents disagree on hosting: the licence allows processing in the EU, the United States or any country where the vendor has facilities, while the privacy policy's transfer register lists only Ireland and the Netherlands.
- Synthetic data that looks convincing is not automatically representative. A model validated only on simulated corner cases can inherit the blind spots of the simulation itself, which is precisely the risk safety validation exists to catch.
- Nothing is published on whether customer material is used to train models, and no opt-out is documented, so raise both in contract negotiation rather than assuming.
- The logos on display are customers, partners and research institutions. Their accreditations and scale say nothing about the vendor's own.
Setup & Integrations
Technical difficulty
Technically light, commercially slow. There is nothing to install: the offering is a web application reached with credentials, and scenarios are configured in the interface without writing code. The real effort sits either side of that. Getting started is not self-service, so expect a demonstration, a scoping conversation and a signed agreement before any access. Then outputs must be aligned to your own taxonomy, ontology and labelling specification so datasets drop into your machine learning pipeline cleanly. The audience is assumed expert: perception, validation and machine learning engineers who already know which sensors and edge cases they need.
Deployment
Supported languages
Behind Anyverse ADAS & AD
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What exactly is Anyverse ADAS & AD?
Which sensors can it simulate?
What comes out of a generation run?
How much does it cost?
Is there a free trial or a free plan?
Does it offer an API?
Who owns the synthetic data that is generated?
Where is the data hosted?
Who is behind the tool?
Is Anyverse certified?
Should you pick Anyverse ADAS & AD?
Anyverse ADAS & AD is a specialist instrument, and it does not pretend otherwise. For a perception or validation team at a car maker or Tier 1 supplier, the case is strong: a spectral render engine built for machine vision rather than for entertainment, sensor models that extend to radar and thermal, fourteen layers of ground truth on every frame, and the ability to conjure the fog, glare and sensor failures that are dangerous or simply impossible to film. The licence terms are notably generous where it matters most, granting perpetual, worldwide and exclusive rights over the generated data, and version 3 goes further by ring-fencing confidential customer inputs. The company is credible, with roots in more than twenty-five years of physics simulation at Next Limit and a TISAX certification obtained in 2026.
The reservations are about transparency rather than capability. Nothing about the commercial offer is public: no price, no tier, no published range, only a form. There is no documented API, which will frustrate teams wanting to drive generation from their own pipelines. And two questions a careful buyer will ask, whether a Data Processing Agreement is available and whether customer material feeds model training, simply have no answer anywhere on the site.
One caution on reading the marketing: Euro NCAP, UNECE, ISO 26262 and SOTIF appear throughout because they are the standards the data helps customers meet, not accreditations Anyverse holds. Treat this as a serious tool to evaluate through a conversation, budget in hand.
- Choosing a selection results in a full page refresh.
- Opens in a new window.