On April 11, 2019, London-based transcription startup Trint successfully raised USD 4.5m in Series A funding in a round led by Horizon Ventures’ seed fund Horizons Lab, with TechNexus and The Associated Press participating. Founded in December 2014 by veteran journalist Jeff Kofman, Trint has raised a total of USD 7.8m to date, according to Crunchbase.
This funding round will go toward enhancing the core AI transcription function of Trint, according to TechCrunch.
In July 2018, Trint received a small amount from Google’s EUR 150m Digital News Innovation (DNI) Fund. The company received EUR 300,000 (USD 339,000) in funding for the Trint Translation Project (TTP), which adds neural machine translation (NMT) capabilities to the company’s transcription platform.
Trint is one of a few emerging AI-based transcription platforms that offer automated transcription services with the capability to post-edit. Tel Aviv-based Verbit, which managed to secure USD 23m in Series A funding in 2018, is another contender.
“Trint uses artificial intelligence to power its web-based automated transcription platform,” Trint Marketing Manager Michael Wolter told Slator in an August 2018 email message. The company’s platform uses automated speech recognition and a frontend that is a cross between text editor and media player.
The platform also supports exporting verified transcripts to various formats and an iPhone app that lets users record, upload, and view transcripts, with future plans to let mobile app users edit transcripts as well.
“Transcription on Trint is currently available in multiple English accents and 12 European languages,” Wolter said. Asked about market demand in the US and abroad, he said it is still very high in the US, especially among large organizations. He added, “We have a strong presence in many European countries and will continue our expansion into these markets as the demand there grows.”
Trint’s clients include media outlets such as BBC, Der Spiegel, The Associated Press, The New York Times, and The Washington Post. TechCrunch noted how Trint has grown from 4 to 45 employees since being established four years ago.
Trint Translation Project
While transcription may be an adjacent market (certainly, AI-based transcription is an adjacent technology), what interests Slator more than the actual platform are plans to make it multilingual.
Wolter confirmed that the TTP will use NMT to let users translate and globally distribute content into over 100 languages. “TTP will rely on neural machine translation to power its multilingual translation. We have no official language partners on board at this time,” he said.
He explained that, for both transcription and translation, Trint gives users a first draft, and the platform’s workflow allows users to “search, verify, edit, share and collaborate with the machine-generated content” — all in keeping with the mandate of their funding source for TTP, the Google DNI Fund, that “helps news organizations adapt to the challenges of journalism in the 21st century with innovative digital technology,” Wolter said.
“Beta versions of TTP will follow the same journey as Trint’s automated transcription tool: if we can create a product that’s useful to newsrooms and journalists, it will be useful for other markets as well,” he said, citing media, government, education, and law as among these markets.