A startup called Webflakes aims to bring some of the best international content on fashion, food, travel, and more to English-speaking readers with the help of volunteer translators. The site is officially launching today, and the company is also announcing that it has raised $3 million in Series A funding.

CEO Nathan Shuchami told me that people searching the web can sometimes struggle to find “genuine, authentic content” on a given topic due to language issues. For example, for wine connoisseurs, there are certainly plenty of sites about wine, but the commentary of many French experts is inaccessible unless you speak French.

To address that issue, Webflakes has selected 60 established bloggers in Japan, France, Italy, Spain, Argentina, Switzerland and Peru and has licensed the rights to their content in every language except the original. Then a team of volunteers translates their work into English and posts it on the Webflakes site.

For now, neither the blogger nor the translator is paid. Shuchami said the blogger gets exposure to a new audience. Meanwhile, many of the translators also do professional translation work, so this is an opportunity to do something more fun and build their portfolio. Plus, Webflakes will donate $1 for every 500 words translated to the charity of the translator’s choice. And in the future, Shuchami suggested that Webflakes might be able to offer revenue-sharing deals to both groups.

But is the global nature of the content enough to attract readers? One advantage, Shuchami said, is that Webflakes is currently focusing on lifestyle topics where the blogger’s nationality should be a particular draw — not just French writing about wine, but also Italian writing about Italian food, Japanese writing about Japanese architecture, and so on.

I poked around the site this morning — I don’t read a lot of lifestyle content, but I thought the range of topics was pretty interesting. The top trending article right now is a French writer on “How To Wear A Bow Tie.” Also on the front page is a Peruvian writer telling readers to “Invite Your Mother To Peru For Mothers Day!” And the translations are usually quite readable, if not always graceful. (To be fair, that may have as much to do with the original post as the translation. And yes, the writing on English-language blogs can be pretty rough, too.)

Eventually, Shuchami said he hopes to add more writers and translators and to expand to other kinds of content.

As for the funding, it was led by Oren Zeev’s Orens Capital, with participation from Genesis Capital, Audible CEO Donald R. Katz, eBay CTO Mark Carges, Chegg co-founder Aayush Phumbhra, former GoDaddy CEO Warren Adelman, former Apax partner Stephen Grabiner, and others.

Article courtesy of TechCrunch

Gengo, an increasingly popular online translation service that uses a network of more than 7,500 pre-screened and rated translators to provide high-quality translations in 33 languages, announced that it has raised a $12 million funding round. The round was led by Intel Capital, with participation from returning investors, including Atomico, Iris Capital, Infocomm, NTT-IP and Saudi Telecom Ventures. The service previously raised a total of $6.8 million, including a $5.25 million series A round led by Atomico and Dave McClure’s 500 Startups. McClure is also a Gengo board member.

The fact that a number of telecom companies are part of this round, Gengo’s CEO and founder Robert Laing told me in an email earlier today, ”shows how telecoms companies ‘get’ the global opportunity of Gengo.”

“The Gengo team is excited about working with investors from Asia, the USA, Europe, and the Middle East, led by Intel Capital, because of their global experience and track record helping entrepreneurs,” Laing writes in today’s announcement.

Added Matthew Romaine, CTO and co-founder of Gengo: “There’s a significant technology component to human translation at scale, so it’s great to work with a firm with the pedigree of Intel Capital.”

Currently, Japan and the U.S. account for about 40 percent of Gengo’s revenue each. The company currently has a staff of 30 in its Tokyo office and nine employees in San Mateo.

According to Laing, the company has been growing rapidly. Gengo’s translators have already translated more texts in 2013 than they did during 2012. Part of this growth, of course, is due to the recent partnership with Google’s YouTube, which has now made Gengo one of its two integrated paid translation services, as well as a recent partnership with 3Play Media.

Besides video, Laing says, Gengo is also seeing a huge volume of translations from travel and e-commerce sites, as well as from a number of “leading e-commerce, online travel, and community portals are powered by its translation platform.”

The Gengo team plans to use this new round of funding to accelerate its global expansion and improve both its translation platform and increase the speed of the translation process.

Article courtesy of TechCrunch

Google Translate is a very useful tool for when you are travelling internationally but sadly, that’s also the time when you are least likely to have an always-on connection to the Internet. Obviously, there are a number of offline translation apps available, but if you are partial to Google Translate and you use an Android phone, you’ll be happy to hear that the latest version of the Google Translate app for Android (2.3+) now lets you download offline language packages for about fifty languages.

You can now simply select [Offline languages] in the app menu and see all the language packs available for download. You just need to download the language packs for the two languages you want to translate between and you are good to go. Google notes that these packs are “less comprehensive than their online equivalents,” but even a smaller dictionary is more useful than not having one at all.

Google also offers a Translate app for iOS, but it’s not clear when (or if) this version will get an offline mode, too.

While the offline mode is obviously the main feature in this new version, the app now also allows you to translate vertical text in Chinese, Japanese and Korean with your camera. Google added support for using camera input to translate texts last August and added basic support for translating Chinese, Japanese and Korean this way last December. This could be a killer feature for Google Glass, too, and it’d be a surprise if the Google Translate team wasn’t working on this already (especially given that Google Translate’s Josh Estelle is already a Glass user and that Translate has made some cameo appearances in Google’s Glass promo videos).

Article courtesy of TechCrunch

Google Translate just added a cool new feature that allows you to easily create a personalized phrasebook with the phrases and sentences you want to memorize and/or find yourself translating repeatedly. As the Google Translate team notes in today’s announcement, the idea here is to allow you to jumpstart the process of committing the translation to memory by “allowing you to save the most useful phrases to you, for easy reference later on, exactly when you need them.”

Revisiting these phrases regularly, Google argues, will help you turn these translations “into lasting knowledge” (just like those rote drills from your Latin classes back in the day).

The new phrasebook is now enabled by default, and you can access it through the little book icon in the top right corner of the Google Translate screen. To save a phrase, simply press the new star icon underneath the translations.

The phrasebook itself is pretty straightforward, with one language on the left and the translation on the right. You can filter phrases by language pairs and – just like across the rest of Google Translate – there is a text-to-speech feature that allows you to listen to each phrase.

Article courtesy of TechCrunch

Duolingo, the popular language-learning service founded by CAPTCHA co-inventor and reCAPTCHA founder Luis von Ahn, just launched the latest version of its mobile app for the iPhone. This new version introduces an offline mode, so users can now do their lessons on the subway and on planes, as well as a voice-recognition feature that tells learners whether they pronounced a word or sentence correctly.

As von Ahn told me earlier this week, offline access has always been one of the most requested features for Duolingo. The app now caches about an hour’s worth of lessons and then syncs with Duolingo’s servers once it is back online. Interestingly, this also means that the app is now always a bit faster, because even when users are online, most of the lessons now don’t have to be downloaded on demand.

The only feature that doesn’t work offline, however, is the new speech-recognition tool because the actual speech recognition is handled by Duolingo’s servers. This new feature, von Ahn told me, checks how close your pronunciation is to a native speaker and then provides some basic feedback.

Duolingo previously offered this feature on its website (and they both use the same backend), but as von Ahn told me, about half of the lessons on the service are now completed on the iPhone. On average, users now spend more time with the mobile app than on the website.

Looking ahead, Duolingo plans to launch an Android app with all the functionality of the iOS app in May.

In addition, the team is working on reimagining the translation experience. As you may recall, the original idea behind Duolingo was twofold and similar to von Ahn’s last startup, reCAPTCHA: Duolingo would teach users how to speak and write a language, and, in return, users would translate online texts into other languages once they got proficient enough to do so (and Duolingo would be able to charge for these translations). Right now, however, von Ahn told me, “users like the lessons more than the translations – and I don’t want that.” So the Duolingo team, which consists of about 26 people, is working on redesigning the translation part of Duolingo, though it’s not clear when the team plans to launch this.

Article courtesy of TechCrunch

Google just launched an update to Google Translate that brings a few new features to the tool that will make it easier to decide which translation of a given word is the one you’re actually looking for.

With this update, Google Translate will now rank possible translations by frequency. Every possible translation is now also marked as common, uncommon or rare and this ranking is based on, as Google says, “the vast number of translations we use to train our system.” Very rare translations will actually be hidden by default, but you can bring them back with a single click.

Google Translate now also groups synonymous translations together when their meaning is closely related. For the time being, this feature is only available in English. Another new feature is reverse translations. Google Translate’s users, the company says, often check the tool’s translations by translating them back into their original language. This, says Google, can help its users “distinguish translations of different meanings and reveal subtle differences among similar words.”

Article courtesy of TechCrunch

A lot of mobile language software wants to be the equivalent of an automatic universal translation app, listening to spoken phrases and translating them using natural language processing and audio recognition. Nice as that is in theory, we’re not there yet. Audio translation isn’t anywhere near perfect, and passing a phone back and forth between two speakers is awkward in practice. Enter Mantaphrase, a startup with a vision of quick, focused, visually assisted conversations across languages that help travelers get exactly the info they need quickly.

Mantaphrase, a Waterloo-based company that released its first iPhone app earlier this month, provides a show-and-tell method for cross-language communication, highlighted by simple design, yes/no answers and slider-based controls for incremental responses. The app organizes interactions by category, and offers follow-up suggestions to keep the conversation flowing. It works offline, and since it’s working from a pre-determined library, while arguably less flexible, you won’t have to worry about miscommunication caused by machine error.

The design of Mantaphrase is simple, but its simplicity belies the fact that there’s a lot of thought behind its user interface. In an interview, Mantaphrase co-founder Wen-Hao Lue explained that the focus was on making sure text presented to the reader was big, clear and easy to spot, and that buttons were large and convenient. They’re still getting it right – he notes that slider size will be made larger in a future iteration – but overall he says the simplicity of design, which was unabashedly influenced by Square, where Lue interned last summer.

“We’ve had testers in China and Japan test this out on people, and we got a lot of good insight into how people use the app,” he said. “The most important thing is large text, and some people don’t understand that.”

At launch, Mantaphrase supports translation between English and Chinese and Japanese. Those languages were chosen based on the founders’ backgrounds (Wen-Hao is originally from Taiwan, and co-founder Patrick Tardif spent 9 months living in Japan), and because good translation sources for stock phrases were readily available. Mantaphrase uses online databases, pares those down to about 200 key phrases, and then categorizes those themselves. The app itself is free, but language packs are $3.99 in-app purchases. Lue says they’ll look at other models going forward, including possible subscription-based pricing.

“Right now, with English, Chinese and Japanese, for North Americans in terms of people traveling outbound, we’re at 10 percent, so 10 percent of travelers outbound go to China, Taiwan, Japan and Singapore, so people where they speak Chinese or Japanese,” Lue said. “We know that if we add Spanish, French, Italian and German we can get to 85 percent.”

To add those languages, Mantaphrase will look for strategic hires, as well as additional online sources of content freely available to use. The company is currently bootstrapped, and Lue says it has been happy with that arrangement so far, but scaling up to address four new languages could require additional outside funds. Also, later on, Mantaphrase hopes to go beyond pure utility and provide tools for users looking not only to communicate, but also to learn additional languages beyond their own. Depending on how Mantaphrase approaches that, its app could become a useful tool in second-language education.

Article courtesy of TechCrunch

YouTube just announced that it is now making it easier for video creators to translate their video captions to more than 300 languages by integrating the YouTube Video Manager with the Google Translator Toolkit. It’s important to note that these are not the automatic caption translations that YouTube currently makes available for about 50 languages. Instead, users who want to support captions in languages currently not supported by the automatic tools (or who prefer hand-crafted translations over the sometimes odd machine translations) will have to do the work themselves or invite others to help them.

As 70 percent of YouTube viewers are now outside of the U.S., says Google, being able to more easily reach international audiences is becoming increasingly important for video creators (but also for YouTube itself) and these tools are meant to help YouTube’s videographers reach this audience.

Because of its integration with the Translator Toolkit, Google’s tool for helping translators collaborate and do their work faster, these manual translations should be relatively easy to create. To get started, users who want to translate their videos need to create a caption track first. After that, they will be able to start translating these captions using the new “request translation” feature in the YouTube Video Manager. For languages supported by the YouTube automatic translation feature, Google will provide a first draft of the translation.

One nice feature here is that users can watch the video in the editor as they work on the translation. This gives the text more context and should make the translator’s life a bit easier.

Article courtesy of TechCrunch

TechCrunch has millions of readers around the globe, and they speak far more languages than our staffers or guests at Disrupt. To make it easier for our global readers to enjoy the conference, we are excited to partner with Babelverse, a Disrupt NYC 2012 Battlefield finalist, to provide real-time, simultaneous translation of the conference (September 10-12, 9am to 6pm PDT).

Anyone watching the live webcast can visit http://tcdisrupt.babelverse.com and select from one of 12 languages and hear a real person (not a language bot!) translate each session.

The 12 languages are:
- Español (Spanish)
- Français (French)
- Português (Portuguese)
- Deutsch (German)
- Italiano (Italian)
- 普通话 (Mandarin Chinese)
- Русский язык (Russian)
- मानक हिन्दी (Hindi)
- 日本語 (Japanese)
- 한국말 (Korean)
- Türkçe (Turkish)
- Arabic

We hope that this is a small step to help startup fans everywhere follow TechCrunch.

Babelverse also recently announced that they raised a seed round led by 500 Startups.

Article courtesy of TechCrunch

For front-end developers, the process of translating a designer’s Photoshop file into a bit of usable code isn’t easy. It requires either breaking down different layers and coding a page by hand, or outsourcing the work out to a conversion shop and hoping for cross-platform capabilities. Either way, the translation typically takes a few days to get done.

Y Combinator-backed Markupwand thinks it has a better way: The startup has created a web application that allows developers to upload Photoshop files and get back well-written, cross-platform code in a matter of minutes.

While Markupwand hopes to reduce the amount of time it takes to hand code or outsource development, it could also replace existing tools like Dreamweaver and Fireworks, which neither developers nor designers like very much, and which tend to do a poor job of creating useable code. So the company puts an emphasis on creating code that looks as good as hand-written code. There’s no absolute positioning and minimal markup and CSS.

It’s kind of a niche application, but could become a life- and time-saver for front-end developers. In a closed beta period, Markupwand has attracted 3,000 developers from 250 companies and freelance firms across 54 countries.

While in beta, Markupwand is making the capability available for free, as it works out the bugs. But it’s planning to eventually charge for the service, either on an a la carte basis for one-off translations, or enabling developers to sign up on a subscription basis. It’s trying to determine pricing now, which could include enterprise licenses for access.

Markupwand was founded by front-end developers Raj Natarajan, Alagu Muthuraman, and Suren Mahendran. Muthuraman and Mahendran new each other from NIT Trichy, where they both graduated. Meanwhile, Natarajan and Muthuraman met at Yahoo! bootcamp after joining the company at the same time. The three founders have worked in various positions at Yahoo, Zynga, Interviewstreet, and Myntra.

Article courtesy of TechCrunch