What Is a Cloud Translation Tool? Advantages and Limits

A cloud translation tool is a translation environment that is not installed on your computer but runs through a browser. The text, the translation memory and the termbase sit on a server rather than on your own device. Translator, editor and project manager can all reach the same file at the same time from different places.

Contents

This model spread quickly over the past decade. Doing away with installation, easier licence management and simpler teamwork were strong reasons. Yet like every convenience it comes at a price: where your data physically sits is no longer under your direct control.

Below we set out what this model means for the buying side, where it is an advantage and where it calls for care.

What is a cloud translation tool?

Technically the difference lies in a single point: where does the software run? With a desktop tool the software runs on your own computer and the files sit on your disk. With a cloud tool it runs on a remote server and all you see is its screen.

The first consequence of this difference is access. You can connect to the same project from the office, from home or from another country. There is no longer any need to send files back and forth; everyone works with a single, current version. Much of the confusion created by numerous file versions circulating by email disappears as a result.

The second consequence is maintenance. Version updates are carried out on the provider’s side. This eases a long-standing difficulty: in one case study, 90 per cent of translators reported having experienced tool compatibility problems (Pietrzak & Kornacki 2021). Having everyone on the same version removes part of that problem.

The third consequence is the cost structure. Instead of a one-off licence, a monthly or annual usage fee is paid. This lowers the barrier to entry for small suppliers; over the long run it may raise total cost. The same case study found that 91 per cent of translators had bought their own licence; the subscription model spreads that burden over time but does not remove it.

Differences between desktop and cloud

The table below compares the two models against the criteria that matter to the buying side. No single row is decisive on its own; the decision follows your document type and your confidentiality needs.

Criterion	Desktop tool	Cloud tool
Installation	Installed on the device, version management required	Opened in a browser
Data location	On the user’s device	On the provider’s server
Teamwork	Through file exchange	Shared access at the same time
Internet	Not required	Mandatory
Cost	Usually a one-off licence	Periodic subscription
Operating system	Mostly Windows	Any system with a browser

The last row is particularly important. The long Windows focus of desktop tools can be explained by that system’s 78.43 per cent share of the desktop market in the July 2019 figures (StatCounter 2019). The cloud model largely removes this dependence.

Advantages for the buying side

The first advantage is speed. When an urgent job comes in, putting a team together takes minutes; giving a new translator access is quicker than sending files. On large, short-notice jobs this has a direct effect on the delivery date.

The second advantage is consistency. Because everyone works against the same memory and the same term list, terminology drift in split texts is reduced. A decision made by one translator appears instantly on the screens of the others.

The third advantage is visibility. The stage the job has reached, how many sections are finished and where the checking stands can be followed in real time. On long projects this makes the delivery date easier to predict.

The fourth advantage can be summed up as continuity. Even if a translator leaves, the memory, the terminology and the project history stay in place. The risk of losing knowledge tied to one person falls, and a new arrival can see earlier decisions from the moment they start.

The fifth advantage is ease of access. Being able to join the project as an observer yourself lets you see interim versions and step in early where needed. On long-running localisation projects this markedly reduces surprises at final delivery.

Limits: connectivity, versions and lock-in

The first limit is dependence on connectivity. An internet outage means production stops outright. On projects with critical delivery dates you should ask whether an offline working option exists.

The second limit is loss of control. When the provider changes the interface or withdraws a feature, the user has no say. With a desktop tool there is the option of staying on an older version; in the cloud there usually is not. An unexpected interface change during a heavy delivery period can slow a team down for a while.

The third limit is supplier lock-in. If your data is stored in the provider’s own format and export is limited, changing platform becomes expensive. Support for standard exchange formats reduces this risk.

The fourth limit is continuity risk. If the provider discontinues the service, what happens to your data has to be secured by contract. This may look like a remote possibility, but it is a matter to be taken seriously for long-lived archives.

Data location and the confidentiality question

In the cloud model the most critical question is this: in which country is the server that physically holds my text? The answer is not a technical detail; it determines which country’s law applies.

Tool makers have registered this sensitivity too. The book reports that memoQ shows a warning dialogue reminding the user of data protection obligations (Pietrzak & Kornacki 2021, p. 25). The very existence of the warning shows that the industry is aware of the issue.

The second point is access rights. In a project running in the cloud, who can see which document can be configured. Setting these permissions correctly keeps confidentiality intact regardless of where the server sits.

The third point is the retention period. How long files will remain in the system once projects are complete, and how they will be deleted, has to be defined. Indefinite retention is a risk that grows quietly.

The fourth point is whether the source text is sent to a machine translation engine. Cloud environments can often be connected to an engine; if that connection is switched on, the text reaches a second provider as well. With confidential documents, confirming that the connection can be switched off is as important as the question of server location.

Clauses to look for in the contract

In a cloud-based arrangement what protects you is not technology but the contract. Having a few clauses in place settles most of the arguments that might arise later before they start.

The first clause is ownership: the translation memory and the termbase belong to you and will be handed over in a standard exchange format on request. Without this clause your past translations effectively stay with the supplier.

The second clause is data location: the country in which the data is processed and whether subcontractors are used must be set out in writing. The third clause is destruction: how long files will be kept after delivery and how they will be deleted must be stated.

The fourth clause is the access list: who may reach the project and whether those people have given confidentiality undertakings must be spelled out. The subcontractor chain is the point at which confidentiality is broken most often.

Is the cloud suitable for every job?

The answer depends on the document type. With documents such as commercial correspondence, marketing copy, user manuals and web content, the cloud model usually works without difficulty. The real gain with these documents is speed and shared access; the confidentiality risk is at a level that can be managed by contract.

Some document types, by contrast, call for a more careful assessment. Files in ongoing litigation, documents relating to merger and acquisition talks, unpublished patent applications and patient records fall into this group. The question here is not whether the technology is adequate, but who carries the risk.

There are two common solutions for work of this kind. The first is to use a set-up that undertakes to keep the data in a particular country. The second is to run the job in a closed environment with a limited team; this is usually slower and more expensive, but it is easier to trace.

The decision criterion can be framed as a simple question: what harm would arise if this document reached a third party without authorisation? If the answer is small, the cloud model can be used comfortably; if it is large, the set-up needs to be discussed separately.

The Türkiye context

In Türkiye the most direct legal counterpart of cloud use is the provisions of Personal Data Protection Law no. 6698 on the transfer of personal data abroad. Documents such as civil registry records, medical reports and criminal record certificates contain personal data; processing such documents on a server abroad requires the conditions set out in the law to be met.

The second point is corporate procurement. Public bodies and companies in regulated sectors may ask their suppliers for a statement of where the data is processed. Having that statement ready at the quotation stage saves time in tender and purchasing processes.

The third point is practical infrastructure. Connection quality is usually adequate in large cities; during heavy delivery periods, however, having a defined offline fallback reduces dependence on a single point.

This article is for general information and is not legal advice.

What does this mean for you?

Whether your supplier works in the cloud or on the desktop is not in itself a good or a bad sign. What matters is that the choice suits your document type and that responsibilities are put in writing. The steps below let you make that check quickly.

1. Sort your documents by level of confidentiality; mark those containing personal data as a separate group.
2. Ask your supplier in which country the data is processed and whether subcontractors are used.
3. Have it written into the contract that the translation memory belongs to you and will be handed over on request.
4. Set the retention and destruction period for files after delivery in writing.
5. For critical deliveries, agree a fallback working plan for connection outages in advance.

To set up a working arrangement that matches the confidentiality level of your documents, you can send us the details through the quotation form, and you can reach our other articles on data security through the knowledge centre.

Frequently asked questions

Are cloud tools less secure than desktop tools?

Not in themselves. What is decisive is the provider’s contract, where the data is stored and how access rights are managed.

Can you work without an internet connection?

In most cloud environments only to a limited extent. Work stops during long outages, so an offline fallback should be agreed for critical deliveries.

If my translation memory stays in the cloud, who owns it?

Ownership is determined by the contract. It needs to be put in writing that the memory belongs to you and will be handed over in a standard format on request.

Is it a problem if my data is processed outside Türkiye?

It is a matter to be assessed for documents containing personal data. The conditions for transfer abroad are set out in legislation and should be checked beforehand.

Sources

• Pietrzak, P. & Kornacki, M. (2021). Using CAT Tools in Freelance Translation. Routledge Focus, p. 25.
• StatCounter (2019). Desktop Operating System Market Share Worldwide, July 2019.
• Personal Data Protection Law no. 6698 (2016), provisions on the transfer of personal data abroad.