The Cyber Resilience Act applies to products with digital elements made available on the EU market. For software, the hardest question often comes first: is what you supply a product, or a service that the CRA does not cover? The answer depends on what the customer actually receives, where the software runs and which remote processing the product needs in order to work. Labels such as SaaS, managed service or subscription do not decide it.
This guide explains the tests, applies them to monitoring software that runs on a client’s infrastructure while the vendor keeps control of it, and shows how to record the decision. In short, the vendor’s control does not by itself take such software out of scope. What matters is whether software is supplied to the client for its use.
Start with what you supply
Three questions come first:
- Is there a product with digital elements? That is a software or hardware product together with its remote data processing solutions. Software and hardware components placed on the market separately are products too.
- Is it made available on the EU market? That means supplied for distribution or use in the EU in the course of a commercial activity, whether for payment or free of charge. The first such supply is the placing on the market.
- Does it have a data connection? Its intended purpose or reasonably foreseeable use must include a direct or indirect, logical or physical data connection to a device or network. A direct internet connection is not needed. The FAQ gives the example of an offline text editor: it does not start any communication itself, but it runs on a host system that does, so it is indirectly connected.
Article 2(1), with the definitions in Article 3(1), (8)-(10), (21) and (22).
Then check the exclusions in Article 2, for example for products covered by the EU medical device regulations or developed exclusively for national security or defense purposes. Scope and product category are separate decisions: our CRA product classification guide covers the category once you know the product is in scope.
Several facts do not settle scope on their own: who owns the source code, whether the customer can see or change it, whether the price is a license, a subscription or a service fee, and what the offer is called. Supply can be commercial even when the software itself is free, for example where the supplier monetizes other services through it.
Software installed on the user’s side
The Commission’s guidance explains when software is a supplied product: it must be provided to a user, obtained by that user and operated on, or as part of, an electronic information system on the user’s side. Software that is downloaded, installed or otherwise supplied to the user and that executes on the user’s system meets these criteria, including browser extensions and applications built with web technologies but supplied for local execution. For the CRA to apply, the software must also be made available on the EU market in the course of a commercial activity and have a qualifying data connection, as described above. Commission guidance, paragraph 20.
In practice this covers desktop programs, mobile apps, command-line tools, local clients for a cloud service, agents and collectors, and software delivered as an image or package for the customer to run on its own systems.
Source code can be a product. The CRA defines software as computer code, which covers both machine code and source code. Whether a supply of code is a placing on the market still depends on the usual test of commercial supply on the EU market. In the guidance’s example, a company licenses source code to a business customer as a product for a customizable internal platform. It places that code on the market even though the customer must adapt and compile it, and it is not responsible for the customer’s later adaptations. By contrast, sharing free and open-source code in a public repository outside a commercial activity, unfinished code shared for testing or review, and sample code in tutorials are generally not a placing on the market. Recital 20 adds that hosting software on an open repository does not by itself make it available on the market. Commission guidance, paragraphs 22-24.
Software that hardware needs is part of the hardware product. The guidance explains that where a device needs software from the same manufacturer to work, such as printer drivers or the companion app of a wearable, the device and that software form one product, even if the software is downloaded separately. Commission guidance, paragraph 25.
Versions and variants. For standalone software, the guidance treats all copies of a version as placed on the market when that version is first supplied. A later version is newly placed only if it is a substantial modification. Builds that differ in their components, configurations or features, such as editions for different operating systems, are separate products (guidance paragraphs 13-16). Our CRA support period guide explains what this means for the support of each version.
SaaS and software used only through a browser
Software that executes remotely and is merely accessed by the user is not, on that basis alone, a product with digital elements. The guidance names web applications and progressive web apps used exclusively through a browser as typical examples. Websites are not products in themselves, even though some of their code runs in the visitor’s browser. Commission guidance, paragraph 21.
Two questions remain for a SaaS offering:
- Does the customer also receive software that runs on its side? A desktop client, a mobile app, a browser extension, an agent or an on-premises connector is supplied software and needs its own scope check, even if the main service runs in the cloud.
- Does the service support the functions of a product? Remote processing is part of a product when the product needs it to perform one of its functions and its software was designed and developed by the product’s manufacturer or under its responsibility, as the next section explains.
The FAQ confirms that standalone SaaS and other cloud solutions designed and developed outside the responsibility of a manufacturer of a product are not themselves products with digital elements. That is a conclusion about the CRA only: it does not decide whether a provider falls under NIS2 as an entity, or what its role is under the GDPR. FAQ section 1.2.
Remote data processing within the product
A product with digital elements includes its remote data processing solutions. Remote data processing is data processing at a distance for which the software is designed and developed by the manufacturer, or under its responsibility, and the absence of which would prevent the product from performing one of its functions. Article 3(2).
The Commission’s guidance reads the definition as three conditions, all of which must be met:
- At a distance: usually outside the user’s environment, in a public or private cloud, at the manufacturer’s premises or at the edge. Where the servers stand and who operates the software do not decide the question.
- A function depends on it: any function counts, not only the core functionality. The guidance lists sending commands to a device, synchronizing files, onboarding, configuration, distributing updates, and identity and access management. A manual alternative does not take the remote function out. Remote analysis of telemetry collected only for statistics or future product development is not remote data processing.
- The manufacturer is responsible for the software: it was built in-house or tailor-made for the manufacturer by a service provider. Software the manufacturer runs on third-party infrastructure or platform services can qualify. A general third-party SaaS application that the manufacturer integrates does not.
Commission guidance, paragraphs 184-200.
Processing that fails one of these conditions is not remote data processing of the product, but it can still affect the product’s security. Where a function depends on a third-party service that the manufacturer did not design or have developed, the guidance asks you to treat it like a third-party component: assess the risks of integrating it, address them in the product itself and exercise proportionate due diligence. Where no function depends on the processing, such as remote analysis of telemetry collected for statistics, consider any risks it introduces in the risk assessment and mitigate them in the product. Commission guidance, paragraphs 192-193 and 201-202.
Where the product ends. The remote data processing that forms part of the product is limited to the software modules responsible for the product’s functions and the interfaces those modules use with external services. Back-end systems that carry out later processing, and with which the product does not interact directly, are not part of it. They remain external dependencies for the risk assessment. The manufacturer’s own IT is also outside the product: human resources, payroll, customer relationship management, build and release pipelines, and security testing such as penetration testing or red teaming. Commission guidance, paragraphs 182 and 205-206.
The guidance also asks manufacturers (paragraph 204) to state in the technical documentation that the product has remote data processing or relies on third-party remote solutions, and to describe them. Our CRA risk assessment guide explains how to assess the whole product, including these parts.
Monitoring software on a client’s infrastructure
We encountered this case in one of our client engagements: proprietary monitoring software that runs on a client’s servers or network while the vendor keeps full control. The client cannot see the code, change the configuration or decide on updates. Does the CRA apply?
Control is not one of the CRA’s scope tests. The tests ask what is supplied, to whom, and where it runs. Neither the client’s lack of control nor the fact that the vendor installs and maintains the software settles the question on its own. The answer depends on which of three arrangements actually applies. The CRA, the guidance and the FAQ do not address this exact combination, so what follows is our reading of the general rules. It is not an official position, and a real case needs its own facts.
The vendor supplies a monitoring application and manages it
The client receives software for its own monitoring needs, the software runs on the client’s systems, and the vendor installs, configures and updates it. This matches the guidance’s description of software provided to the user and operated on the user’s side. Vendor-only administration does not change that: a managed product is still a product. On these facts, we would treat the software as within scope, provided it is made available on the EU market in the course of a commercial activity, has a qualifying data connection and no exclusion in Article 2 applies. If the vendor develops the software, or has it developed, and markets it under its own name or trademark, the vendor is its manufacturer under Article 3(13). Otherwise, identify who is. The vendor’s management role can then help it meet obligations such as distributing security updates.
The vendor uses its own tool to deliver a service
The client buys a monitoring service: alerts, reports or analysis from the vendor’s specialists. To deliver it, the vendor deploys its own tool in the client’s environment, but the client receives no software for its own use. Products manufactured for one’s own use are not placed on the market, and the FAQ gives the example of development and configuration tools that a manufacturer develops for its own use, unless they are also placed on the market as separate products. If no software is supplied to the client, the CRA’s market trigger may not be met. FAQ section 1.5.
This conclusion needs evidence, because the tool still runs on the client’s systems. Weigh the contract and the actual set-up together: whether the client has any right to use the software or its functions, whether the client’s staff actually use it or only receive the vendor’s output, and what happens to the installation when the service ends. No single fact decides it. Supply can be temporary: the Blue Guide, section 2.2 treats hire and leasing as forms of supply, so removing the software at the end of the contract does not on its own show that nothing was supplied. Calling the offer a managed service is not enough, and software built for a single client can still be made available on the market. Even outside the CRA, the tool’s security matters to the client, which has to manage the risks of what runs on its systems.
A local agent with a vendor back end
Many monitoring products combine a collector or agent on the client’s systems with analytics, dashboards or remote management on the vendor’s platform. Check the agent first, as supplied software. If it is a product, test each remote function against all three conditions above. Where the agent cannot be configured, updated or deliver its alerts without the vendor’s platform, and the platform software that performs those functions was designed and developed by the vendor or under its responsibility, those modules and their interfaces are likely part of the product, even if a third party hosts or operates them. Later processing with which the agent does not interact directly, and third-party services the vendor did not design or have developed, remain dependencies for the risk assessment.
What “no control” can mean
Before applying any of these arrangements, pin down what “no control” means in the case at hand. It can mean that the client has:
- no access to the source code;
- no configuration or administration rights;
- no say over when updates are applied;
- no right to start, stop or remove the software;
- no right to use the software at all, only to receive the vendor’s output.
These facts do not weigh the same. Restrictions on code, configuration, updates or removal describe how a supplied product is managed, and do not by themselves take it out of scope. Whether the client has any right to use the software, and whether it actually does, matters more for telling a managed product apart from the vendor’s own service tool, together with the other evidence above. A client that owns the server but cannot administer the software can still be the user to whom the software is supplied.
Record the scope decision
Keep the reasoning together with the evidence. For most portfolios, a short record for each offering is enough:
| Question | What to record |
|---|---|
| What is delivered | Executable, image or package, agent, appliance, source code, dashboard access, analysis or reports |
| Who supplies it, to whom and where | Legal entity, customer and market; wording of the offer and the contract |
| Where each part runs | Customer systems, vendor platform, third-party services |
| Customer rights | Use, configuration, updates, removal, and what the customer keeps when the contract ends |
| Remote functions | Each function that stops without a remote module, and who designed that module |
| Dependencies | Third-party services and back-end systems outside the product, with their risks |
| Conclusion and open points | In scope, out of scope or pending; the reasoning; what still needs confirming |
This record is an example of implementation based on CRA Articles 2 and 3 and the Commission’s guidance and FAQ. Its fields and format are not mandated by the CRA. For products in scope, the risk assessment and technical documentation required by Article 13 and Annex VII remain binding.
Review the record when the delivery model changes, for example when you add an on-premises edition of a SaaS product, introduce a new agent or move a remote function to a third-party service.
If the software is in scope
Scope is the first step. For an in-scope product:
- Classify it: the category decides the conformity assessment route. Our CRA product classification guide explains how.
- Assess the risks of the whole product, including its remote data processing and the dependencies around it.
- Check the dates: the main obligations apply to products placed on the market from December 11, 2027. A product placed on the market before that date is covered only if it is substantially modified on or after December 11, 2027. Reporting under Article 14 has applied since September 11, 2026 under Article 71(2), and Article 69(3) extends it to in-scope products placed on the market before December 11, 2027. Article 69.
How Meroi Security can help
Meroi Security can assess the CRA scope of your software offerings one by one, map each product’s boundary and remote data processing, and document the reasoning and evidence for each decision. Explore our Cyber Resilience Act compliance services.
In the regulation
Related guides
- EU Cyber Resilience Act Product Classification: Categories and Conformity Routes Cyber Resilience Act product classification: Annex III and IV categories, core functionality, conformity routes and when a notified body is needed.
- Cyber Resilience Act Cybersecurity Risk Assessment: Article 13 and Annex I in Practice How to carry out the CRA cybersecurity risk assessment: scope, intended and foreseeable use, Annex I applicability, residual risk, components and updates.
- Cyber Resilience Act Support Period Requirements Determine and document a Cyber Resilience Act support period, including product lifetime, supplier dependencies, software versions and update availability.
Sources
- Regulation (EU) 2024/2847 (Cyber Resilience Act)
- Commission guidance on the application of the Cyber Resilience Act, OJ C/2026/6013, November 17, 2026, sections 2 and 8
- Commission services CRA FAQ v1.4, September 4, 2026, sections 1.2, 1.3 and 1.5
- The Blue Guide on the implementation of EU product rules 2022, section 2.2