ANDATA Software
ANDATA Software is a modular suite of MATLAB toolboxes for engineering teams: it covers the full chain from raw sensor signals to validated, exportable machine learning models, and adds stochastic scenario simulation for vehicle safety, automated driving and robotics.
What is ANDATA Software?
ANDATA Software is a modular development environment for analyzing large volumes of data, typically sensor signals and multi-source time series, and for building, applying and safeguarding machine learning and mathematical models. The publisher describes it as covering the whole process chain for the application and safeguarding of machine learning procedures and the development of artificial intelligence, including control and decision-support algorithms.
Four core programs make up the suite. Stipulator handles signal composition and analysis: reference signal search, normalization, plausibility checking, partly automated correction of faulty data, sensor and disturbance models, attribute computation, heuristic labeling, numerical conflict analysis, distance metrics and export towards the next stage. Brainer is a unified graphical environment for neural networks, support vector machines, classification and regression trees and classical regressions, with meta-modeling, committees of models, automatic variable selection, robustness evaluation and direct export of trained models to C, FORTRAN, Python, m-Code and XML. Expectator performs automated plausibilization and anomaly detection; it explicitly refuses the single-algorithm approach and combines simple criteria into committees to cut false positives, using self-organizing maps and frequency analysis, and it separates a generic offline tool from specialized online embedded variants. SceneInspector implements incremental probabilistic simulation, multi-stage stochastic simulation and partly automated scenario generation, and manages simulations distributed over a cluster while driving third-party engines such as PC-Crash, VISSIM, SUMO and CarMaker. Extension toolboxes add Signal Structure, Geometric Operations, FEM Operations and Geographic Info capabilities.
Technically all tools are cooperating, complementing MATLAB toolboxes that can also be executed as standalone software, and ANDATA belongs to the MathWorks Third Party Connections Program. An apps and toolboxes concept layers a generic core, customer-specific toolboxes and tailored apps so that users get quick access without having to master everything. The current release is 6.1, covering MATLAB R2023b to R2026a and tested on Windows 11. A further tool, DataDiver, has been announced in a first beta reserved for selected customers, for deep comparison of large numerical simulation results, dimension reduction, clustering, outlier detection, parameter importance and sensitivity and robustness analysis.
Typical fields are vehicle safety, mobility and traffic, connected and automated driving, robotics, control of complex systems, time series prediction and virtual product development. The publisher is an Austrian engineering and deep tech company based in Hallein near Salzburg with a Vienna office, describing itself as family-owned, and it runs an ISO 9001:2015 quality management system certified by TÜV Süd, plus TISAX participation.
What it does
- Collect, clean and normalize sensor signals coming from heterogeneous sources with Stipulator
- Label signals and formulate the classification or prediction requirements a model must meet
- Train, validate and score machine learning models through the Brainer graphical interface
- Detect anomalies and incidents in time series with Expectator
- Generate and run multi-level stochastic scenario simulations with SceneInspector
- Export trained models to C, FORTRAN, Python, m-Code or XML files
- Analyze finite element simulation results with the FEM Operations Toolbox
When to use ANDATA Software / When not to
A quick filter to help you decide if ANDATA Software is the right fit.
When to use ANDATA Software
- Vehicle safety and automated driving engineers who must develop machine learning functions and safeguard them at the same time
- Teams whose job is to validate, stress-test and document a model, not merely to train it
- Organizations already standardized on MATLAB and on the MathWorks toolboxes the suite builds upon
- Engineers working on heterogeneous sensor signals and multi-source time series coming from test benches, simulation and artificially generated data
- Simulation specialists doing scenario management, cluster-distributed stochastic simulation and post-processing of finite element crash results
When not to use ANDATA Software
- Teams without an available MATLAB license, which the vendor states is the basic requirement for using the tools
- Buyers who need a published price list or a self-service purchase, since every quotation goes through the vendor's sales address
- Anyone expecting a web, mobile or cloud product, or a public API to integrate with: none of these is offered
- Users who want to evaluate the software on their own before talking to anyone, as no free plan is announced
- Non-Windows environments, and business users looking for a generalist AI assistant rather than an engineering workbench
How to use ANDATA Software
A typical end-to-end flow, from setup to results.
- Check the prerequisites first: an available MATLAB license is the basic requirement, plus the MathWorks toolboxes each tool depends on
- Add the Statistics and Machine Learning Toolbox for Stipulator, the Deep Learning Toolbox on top of it for Brainer, and note that the Parallel Computing Toolbox is beneficial throughout
- Plan for an external simulation package as well if SceneInspector is part of the scope
- Request a quotation and the licensing scope that fits your team from sales@andata.at
- Supply the technical identifier the license will be bound to: the user name returned by whoami, or the machine MAC address read with ipconfig /all on Windows or ifconfig on Linux
- Give the name of the organizational unit instead when applying for an SME or an industry license
- Sign in to the website login area to download the software packages and extensions
- Install the tools into your MATLAB environment and work through the manuals, which are integrated directly into the MATLAB help pages
- Book training if needed: introduction courses to the tools, expert courses on Stipulator, Brainer or Expectator, methods courses, or tailored inhouse courses via training@andata.at
- Raise issues through the ANDATA Software Support form, which prefills automatically when opened from the Help and Support menu of the tools, by email to support@andata.at, or by phone
Pros & Cons
Pros
- Covers the whole chain, from raw signals to a validated model exported towards an embedded target
- Puts explicit emphasis on validation and safeguarding rather than on training alone, through conflict analysis, robustness evaluation and committees of models
- Builds on an established MATLAB foundation instead of an opaque proprietary stack, the publisher being a member of the MathWorks Third Party Connections Program
- Open architecture: model types, training procedures and scoring functions are customizable
- Runs on the customer's own machines, so the data being analyzed never leaves their infrastructure
- A perpetual license is available, letting a customer keep the purchased version indefinitely
- Quality management certified ISO 9001:2015 by TÜV Süd with TISAX participation, backed by documented methodological depth: around thirty technical FAQ answers, an active software blog and expert training courses
Cons
- No public pricing at all: the entire grid is quotation-only through sales@andata.at
- Strong dependency on MATLAB and on additional paid MathWorks toolboxes, which the buyer must budget separately
- No free plan is announced, and nothing on the site lets a prospect evaluate the tools independently
- No API, no SDK and no developer documentation is published anywhere on the site
- Tested platforms are limited to Windows 10 and 11, and there is no web or mobile application
- No terms and conditions are published, so contractual conditions cannot be read before contacting the vendor, and no data processing agreement, sub-processor list or hosting country is disclosed either
- Breadth that the publisher itself admits takes a while to master, and SceneInspector additionally requires third-party simulation software
Pricing & Plans
No amount is published. ANDATA displays no price anywhere on its website and states that financial offers and pricing can be requested from sales@andata.at, which makes this a quotation-only product. No free plan is announced either, the licensing page listing three paid license types and mentioning no free tier. What the publisher does document is the structure rather than the figures. Licenses are granted on four scopes: user based, machine based, SME for organizational units of up to 250 users, and industry for whole companies. Three license types apply to those scopes. A perpetual license gives unlimited use in time of the purchased version and includes maintenance and support for one year. A maintenance license provides upgrades to the most recent versions during its period, requires a perpetual license to have been purchased beforehand, and can be extended annually. A subscription is a rental limited to the subscribed duration with no prior perpetual purchase, and allows the user or the machine it is bound to be changed once a year. Tools can be bought individually or in packages at reduced bundle prices, again without any figure being stated. The licensing options were modified from 2016 onward. Prospective buyers should therefore expect a commercial conversation before any budget can be estimated, and should add the cost of the MATLAB licenses and MathWorks toolboxes the tools require.
Data, GDPR & hosting
A consolidated view of how ANDATA Software handles your data.
GDPR overview
GDPR implementation is stated but lightly documented. The privacy declaration says data is processed only on the basis of the statutory provisions, naming the GDPR and the Austrian TKG 2003, and it lists the rights available to data subjects: information, rectification, deletion, restriction, transfer, revocation and protest. Complaints can be addressed to ANDATA at dataprotection@andata.at or to the national data protection authority. ANDATA also presents itself as a data mining company aware of the rigor that customer data demands. Beyond that the public record is thin: no data protection officer is named, no data processing agreement is offered, and no sub-processor list or hosting country is published. As the publisher is an Austrian company operating from within the European Union, an Article 27 EU representative is not required, so its absence is normal rather than a gap.
Who owns the data?
The tools are installed and executed on the customer's own workstation or cluster, so the data analyzed with them stays inside the customer's infrastructure and never has to reach the publisher. Ownership itself, however, is not documented in public: no terms and conditions are published anywhere on the website, and the privacy declaration covers only website visitors and the accounts of the authenticated download area, which serves the authenticated access of authorized users to download the software packages and extensions. Nothing published states who owns customer analysis data, what the publisher may do with it, or with whom it could be shared. Those terms only become readable after a commercial contact.
Reuse rights
No terms and conditions are published, so no reuse permission is either granted or restricted in writing. What the site does document concerns the website rather than the software. Account data serves only access to the download area and to protected content, and the behavior of authenticated visitors is occasionally inspected to detect illegal intrusion. Contact and support requests are emailed to the responsible people at ANDATA and manually entered into a ticketing system, which a customer may object to at the start of the exchange. IP addresses attached to blog and news comments are stored internally against abusive use. Web analytics runs on a self-hosted Matomo instance with anonymized IP addresses, and ANDATA states that all analysis data stays on its own servers and is not passed on to third parties. There is no newsletter and no direct automated advertising, though Google maps embedded in some contact forms allow Google to analyze behavior there. For the data processed by the software itself, the tools run locally and the site says nothing about the publisher accessing it.
Data retention & training
Hosting summary
No hosting country and no hosting region is published for product data. In practice the question is largely displaced by the deployment model: the tools are installed and executed on the customer's own workstation or cluster, so the data being analyzed does not have to be transferred to the publisher at all and stays under the customer's own jurisdiction and controls. The only hosting arrangement documented on the site concerns the website itself. ANDATA states that it operates its Matomo web analytics exclusively on its own web servers, so all analysis data remains on those servers and is not passed on to third parties. The authenticated login area is used to download software packages and extensions, not to process customer data. Nothing is published about a data processing agreement, a sub-processor list or the jurisdictions any publisher-side system would fall under, so a buyer with hosting constraints has to raise the question directly with the publisher, an Austrian company.
Things to keep in mind
Risks and trade-offs to weigh before adopting ANDATA Software.
- Budget the total cost rather than the ANDATA license alone: MATLAB and the required MathWorks toolboxes are separate purchases that can dominate the bill
- No terms and conditions are published, so contractual conditions only become readable after a commercial contact, which is late in a buying process
- User-based and machine-based licenses are tightly bound to an account name or a MAC address, and under subscription the user or the machine may be changed only once a year
- A maintenance license presupposes that a perpetual license was purchased beforehand, which changes how a multi-year plan has to be built
- Tested platforms are limited to Windows 10 and 11: check your environment before committing
- SceneInspector assumes you also acquire a third-party simulation engine, and DataDiver is only a first beta reserved for selected customers, so neither should be counted on lightly
- No data processing agreement and no sub-processor list is published; and because the tools automate plausibility checking and labeling, treat their verdicts as engineering evidence to be reviewed, not as a substitute for the reviewer's own judgment
Setup & Integrations
Technical difficulty
High for a newcomer, moderate for a team already running MATLAB. Installation presupposes a properly licensed MATLAB environment, and the license request itself needs a precise technical identifier: the account name returned by whoami, or the machine MAC address. The apps and toolboxes concept exists explicitly to lower the entry barrier for non-experts, manuals are integrated into the MATLAB help, and introduction and expert courses are offered. Even so, the publisher admits the function spectrum is extremely broad and that becoming an overall expert takes a while. The intended audience is engineers and data scientists, not business users.
Deployment
Integrations
Behind ANDATA Software
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Do I need a MATLAB license to use ANDATA Software?
How much does ANDATA Software cost?
Which licensing options are available?
Is a free version available?
Which operating systems and MATLAB versions are supported?
Which tools make up the suite?
Is there an API?
How do I get support?
Are training courses available?
Who publishes the software?
Should you pick ANDATA Software?
ANDATA Software is a specialized engineering workbench rather than a mainstream SaaS product, and it should be judged on that basis. Its value lies in covering the complete machine learning chain, from signal preparation through training, scoring and anomaly detection to stochastic scenario simulation, with an unusual emphasis on validating and safeguarding the resulting models rather than simply producing them. Conflict analysis, robustness evaluation, committees of criteria and export to C, FORTRAN, Python, m-Code and XML make it a credible option for teams that develop safety-relevant functions and must be able to demonstrate that those functions hold up.
Two filters decide whether it is worth pursuing. The first is MATLAB: an available license is the basic requirement, and each tool adds MathWorks toolboxes of its own, so the real budget is the sum of several licenses. The second is the complete absence of published pricing. Every figure comes from a quotation, no free plan is announced, and the missing terms and conditions mean the contract itself only becomes readable through the same commercial conversation.
On the reassuring side, the publisher is an Austrian engineering company claiming a long track record in applied artificial intelligence, running an ISO 9001:2015 quality system certified by TÜV Süd with TISAX participation, and documenting its methods openly through detailed technical answers, release notes and training courses. The software runs on the customer's own machines, so the data being analyzed stays in-house.
In short, this is a tool for engineering and data science teams already living inside MATLAB who need methodological depth and evidence, not for organizations looking for a quick, self-service AI platform. Anyone considering it should budget the surrounding licenses, confirm the Windows requirement and expect a serious learning curve.
- Choosing a selection results in a full page refresh.
- Opens in a new window.