One-tap video app Trash wins NY:LON Connect startup showcase


One of the traditional highlights of the NY:LON Connect conference is the startup pitches session, and 2020’s event was no exception.

New York startup Trash emerged as the winner in an audience vote for the best pitch, but fellow pitchers Super Hi-Fi, MyPart, Boomy and FeedForward won plenty of plaudits from attendees too.

Music Ally first wrote about Trash in June 2019, when its app first launched. The company’s founder is Hannah Donovan, formerly GM of Vine, and before that, co-founder of music-discovery site This Is My Jam.

Trash is a ‘one-tap video app’ that uses AI technology to help people create videos with soundtracks from the material they’ve shot with their smartphones.

“In the last decade we’ve become pretty good amateur photographers… and low-key pretty good models! Today we’re in a new renaissance, and that is video. Anyone can shoot… but putting these moments together is still really difficult. Editing is really tedious, and it takes forever,” said Donovan in her pitch.

She described Trash as “Polaroid, but for video”. Users choose clips from their phone, and the app edits them together. The resulting videos can be shared to the Trash community, as well as to external social apps and networks.

Donovan also explained how Trash works with music. The app has a library of music that people can use to soundtrack their videos.

“Sound is absolutely critical to video. And this is a huge opportunity for music,” she said. “We’ve licensed music from a variety of sources, including production music, working with independents directly. Now we’re starting to talk to larger distributors because we have this non-stop demand from our user base for popular music.”

Trash may have won the audience vote, but the other four startups also impressed with their pitches.

super hi-fi

Super Hi-Fi is a company Music Ally profiled in August last year, with its technology using AI to automatically fill the gap between songs on streaming services with relevant ads, interview clips, artist announcements and radio-style audio branding.

“We think that digital services have an opportunity to be just like the music they put out: dramatic, energetic… and create really unique emotional relationships with their listeners that they don’t currently have,” said CEO Zack Zalon.

“Today, the space between the songs is really a wasteland of dead air. There’s a song, then a gap of silence, then another song.”

Super Hi-Fi is describing its technology as ‘computational music presentation’ – “an extremely powerful AI that can make real-time production decisions that sound like somebody sitting in the radio studio… and it’s a neural network that gets better over time.”


Israeli startup MyPart popped up on Music Ally’s radar in September 2019, when it joined the Abbey Road Red accelerator in London. Its technology uses AI for ‘song mining’ – deep analysis of tracks’ music and lyrics.

“We can turn any type of music catalogue into a song mine,” said CEO Matan Kollnescher. The tech can analyse songs, identify their musical characteristics and lyrical themes, not to mention all manner of rhyming structures.

Clients, for example labels using MyPart for sync or A&R purposes, use ‘reference songs’ to search their catalogues, finding tracks that have similar properties. ““We can tell you not just that this is a Christmas song, but aesthetically it’s a fantastic fit for what you’re looking for,” being one way Kollnescher described it.

He also revealed a future feature aimed at songwriters: “A tool based on the same technology that crunches your song, that can tell a songwriter how close his existing song is to what he wants to achieve”. So they would pick a set of reference songs they’re aspiring to, and the tech “will tell you this is not repetitive enough, the lyrics are not sexy enough, this is too introspective… that’s super useful as a songwriter.”

boomy alex nylon

Music Ally readers should know all about Boomy: we profiled the company in July 2019 (and according to CEO Alex Mitchell, our original news piece came out before he’d even told his mother about the new venture). Its focus is on helping people to ‘make instant music’ using its artificial intelligence system – and even release that music onto streaming services.

At NY:LON Connect, Mitchell played some of the latest examples, and explained how the company owns the copyrights to music created using its AI, but that people can buy that ownership for between $5 and $20.

It also licenses and sells tracks to production music libraries, with a third business strand involving helping people to make money from their online videos through the music.

“It’s really hitting on TikTok. People are making this music to go with their videos, and when they get streamed, be it on Spotify or wherever, we have a rev-share and we pass that royalty back to the user,” he said.

“People are using this music and streaming it not because it’s better than what’s coming out of the industry,” he added. They’re streaming it because it’s theirs, or because they found it in a video they loved.

There is scale building here too. “In 2019, our users created over 310,000 original songs, and in the past couple of weeks we’ve probably seen 20,000 or 30,000 additional,” he said.


Finally, British startup FeedForward showcased its own use for artificial intelligence technology. CEO Lydia Gregory explained the role that its team including mathematicians and software developers aims to play.

“We are the music industry’s AI department,” she said. “We want to perfect the art of turning the latest machine-learning research into useful products for the music industry today”.

Those products include Figaro, an AI search engine for music and sound catalogues, which uses the latest deep-learning technology to find tracks, and manage metadata.

Figaro can find related tracks from any track or playlist within a catalogue; can search a catalogue using an audio track; can automatically tag tracks using a client’s own taxonomy; and manage their metadata from a dashboard.

In tagging, it will suggest tags for racks based on keywords, genres, and instrumentation. “When they get large amounts of new music, Figaro is able to apply tags from their taxonomy with the meaning embedded in that taxonomy to those tracks,” explained Gregory.

“We see this need for intuitive search as being of such importance to every aspect of the industry, to enable us to cope with voice interfaces, to do more sync deals, and to deal with personalised advertising. At the moment, the music industry is not set up to do that, and that’s the problem we’re solving with Figaro.”

Read Music Ally’s other NY:LON Connect coverage
Tech, innovation and music: ‘Pick one little thing and go deep!’
Troy Carter talks music disruption: ‘I’d still be in a panic if I were a major label…’
The evolution of labels: ‘What’s your value proposition?’
Global music streaming: ‘We are just on the tip of the iceberg of what this will become’
Publishing and rights: ‘Everybody needs to play ball in order for this tech to work’

Written by: Stuart Dredge