• Translate a broad range of content into multiple languages in seconds with AI
• Source qualified linguists on-demand through Smartcat's AI-powered Marketplace
• Save hundreds of hours of effort through AI human automated workflows, payment processes, and QA.

Q: What is the reason behind the cost difference in the post editor choices?

A: It's because translators can set their rates. So different translators will have different rates. This gives you more budget flexibility when contracting freelancers.

Q: Does Smartcat utilize Deepl only in the technology behind translations, or are there additional sources that feed the AI translation?

A: Smartcat works with many AI translation providers. DeepL is one of them but we have an algorithm that chooses the best engine for each language.

Q: About the mp4 file you put in the project, what is the output?

A: You can export VTT and SRT files. And even an MP4 with burned-in subtitles.

Q: For Chiara Zannoni of Babbel, do you still use your freelancers outside of the Smartcat marketplace? If so, are they operating and invoicing through Smartcat currently?

A: Yes, we still work with our freelancers that we onboarded to the tool sort of platform last year, when we migrated to Smartcat, and they figure as part of our team within Babel, and the way it works now is that they became Smartcat’s contractors as opposed to Babel. So they get paid directly from Smartcat.

Q: Can users add their glossary to enhance the AI translation accuracy?

A: Yes, Smartcat has a terminology module where you can import your terminology and it can be used for terminology replacement with some AI engines.

Q: Can I work with InDesign files? If so, taking into consideration it's a considerably complex file format, are there any changes I need to manually make?

A: Yes, you can work within design files in Smartcat. The only requirement is that in InDesign you save the document as idml, which is an InDesign markup language from Adobe. So you take this idml file into Smartcat. Then the formatting will be preserved. So when you bring back this idml that's been translated back into InDesign, your fonts will be preserved, your links to your graphics will be preserved, and also all this will be maintained in addition. By the way, we also have a script to help you work with other Adobe tools like a photoshopping illustrator. We have scripts that extract the text from your Photoshop or your illustrated documents. You transmit them into Smartcat and put them back into those applications, and it's going to save you a lot of time rather than copying and pasting texts from this application into, for example, a Word file with a text file.

Q: A given project created in Smartcat may include different files. Is it possible for a linguist in the marketplace to accept just one file depending on the size and reject the other files?

A: You can split assignments between different translators, that's an option. So, for example, if you have 4 documents in one assignment, you could choose to assign one document to a different linguist. For example, if you assign 4 files to a linguist, if they could accept just one of them and not accept all 4. So you would just accept all 4 files or decline the task, but you can communicate with your project manager directly on the platform. 

Q: We have a list of several words that machine translation consistently translates in the wrong context. Could you explain more about how the AI edit learning works?

A: When you click those little checkmarks when a change is made, it's gonna store that in your translation memory. So as you build up that translation memory, some of our customers will come to Smartcat already having a translation memory that they can upload to Smartcat. When you run something through that machine translation, it also pulls from that translation memory and translates it properly. From there you can also upload glossaries into Smartcat as well with specific terminology or acronyms. Things like that that you want to translate a specific way. So that's how that part would work inside the editor.

Q: If so, when using typesetting, how could I work with it when translating to a different alphabet?

A: In this case, what you would need to do is simply modify your style sheets inside the InDesign and replace the font.

Q: Following up on the question above - does it mean the Smartcat will automatically extract the subtitles from the video?

A: We have automatic transcription, so you don't need to upload the subtitles with the video, you can just upload your straight MP4. And then Smartcat will transcribe the text, the speech from the video into something editable. You can even edit the output of the transcription. And it supports many languages, not just English. So if you have a video in different languages, it can be uploaded into Smartcat transcribed, and then, of course, proceed to translation. Then we give you the option to edit to source an edit target, and maybe we can preview a little bit, we're gonna have a much better editor for working with videos in the coming weeks. That's gonna give people a lot more flexibility to edit the output of the transcription.

Q: Are you offering/planning to offer also real-time AI interpretation service?

A: So that's not a real-time interpretation service that we offer. But moving forward, we're gonna offer the option that when you translate an MP4 video, you'll be able to transcribe the video to the text, and then we'll give you the option to generate a voiceover automatically, using, for example, a different voice. And then well, you have the option to pick a voice. Pick the tone of voice and things of that nature. So it's not really like interpretation, but that's one step forward to having the full workflow of translating a video from transcribing the video translating video and then generating new voices in different languages.

Q: Is there a way to import reference material automatically through i.e. GitHub store?

A: With some file formats. Also, you can add links to pictures, and when you import the documents in Smartcat, we can read those links and then show you a preview, for example, for specific segments where people would translate product descriptions from their e-commerce websites, and they send us like a file format. And I think it's po file format, where you can have comments, and then in the comments, you could have a link to a picture. So someone translates the description for a dress or t-shirt, or whatever it is, and then, when the transit is inside this market editor, they can have a picture of this article that they are describing and translating

Q: What is the security like? Does our data get shared with other engines?

A: First of all, you know, Smartcat is hosted on the Amazon AWS platform, which provides, of course, security. We also have SOC type 2 compliance. We are also GDPR compliant. So lots of compliances there regarding the security of your data. And, second of all, when you submit documents on Smartcat, you also have the option to prevent the translators from accessing the documents that are being translated. They only work online. This is one of the benefits of a platform like Smartcat, which is cloud-based when there is no file exchange needed. Everybody works online. And of course, when you have a workspace, you can change the access right? So only people who have specific access right to go to your workspace will be able to see your documents, create projects, and download any of the files that you have, so we feel pretty confident that we have a high-security level on Smartcat. Also, all the freelancers have to sign our terms of service, so your confidential information is secure with our freelancers from the marketplace.

Q: Do you have an analysis of best to worst languages supported by AI translations?

A: What we implementing right now in Smartcat which could be helpful is that we have what's called TQS - Transition quality score, when you run a document through a Smartcat AI when the document has been translated, we're gonna give you a score from 0 to 100, which is gonna give you an idea of what our system thinks about the AI transition. How good it is, and then we give you an average for the document. And then, when you open the document inside the editor that Daniel was showing you, you'll be able to even see individual scores for each of the sentences which is, gonna allow you to focus just your efforts on, for example, on sentences which may be having a lowest core. So your document score might be 90, or 85 could be 95. But then some sentences might be 70, 75, 80. So rather than spending time on sentences that score very high, you can spend more time on sentences that have a low score and have a more homogeneous quality inside the document.

Q: We currently use Phrase, which offers webhooks for different instances (e.g., job workflow statuses/running analyses). We use these webhooks to integrate with our internal systems — does Smartcat also offer similar webhooks for enterprise teams and how robust are they?

A: We have a full set of API methods. We even have a common line interface, that you can use. So there's everything you can do. You can do integrations with your business applications, using smartcat. And it's quite robust. We've done some integrations for a Babble, so they can hopefully testify to the robustness of the integrations. And the difference that we have is a marketplace. That's not something that our competitors don't have. But we have an industry first, AI sourcing, and an embedded marketplace, which is one of the ways that we sign apart. And because of that we understood the use case and that people need to source linguists and work with them at the same time, without having to use multiple platforms for that purpose.

Q: As a translator, can I upload my previous translations to the translation memory of the platform?

A: Yes. You can import TMX files.

Q: (For Chiara) Based on your experience with Smartcat, how has the quality been for “creative” text (e.g., literary text, catchphrases, memes) using the AI translation > editing workflow?

A: So to provide a bit of context on how we've been using it and testing it out. I think our magazine tests are a good example of this, we're currently testing out if using the Openai Gpt Mission Transition engine that Jean Luc mentioned earlier. That gives you the possibility to create prompts as well to guide the machine translation can enable us to instantly translate magazine articles while ensuring that cultural references are converted. And what we found is that that work seems to work. Well, for straightforward articles that do not include a lot of idiomatic expressions, for example. So this is what we've we've been testing it out for mostly so articles that include a lot of idiomatic expressions and very specific cultural references are still a bit harder to to localize using the AI functionalities we found.