Songclip had the idea of building an extensive library of music clips and allowing social and digital companies to offer it as a feature. Like Giphy, the company came with a new media format: the short audio clip. STRV was called to make it a reality.
First, we had to create a one-of-a-kind messaging platform that would put the clips to use. This required an iOS app that could generate thousands of clips automatically, with minimal user input and across multiple platforms. Then, a few years later, our Data Science team jumped in to help with the API platform itself by automating manual processes like music tagging and lyric extraction.
Back in 2016, the initial challenge was the app itself. This project required a tech stack that could take a full-length song, cut it into small pieces and assign an image, sound or video to it. The resulting content would then be available in the mobile app immediately. Songclip’s agreement with Sony also meant that all available songs had to be downloaded and usage closely tracked, so we’d need a tool for data ingestion and reporting.
Fast-forward to the challenge our Data Science team took on in 2021. The API platform had millions of songs in need of being manually analyzed, tagged and matched with exact lyrics, with its database expanding daily — impossible via manpower alone, but a matter of minutes with the right AI solution.
The STRV Solution
Setting up the tech stack involved developing production-ready backend APIs that covered the consumer app (iOS) and the admin interface where Songclip’s content team could manage the songs, videos and lyrics presented to the user.
With that done, Songclip hit the ground running, investments flooded in and the goal became for music clips as a feature to become the norm. STRV’s Data Science team came in to build the multiple AI solutions needed to get there.
As of today, we’ve delivered AI solutions that classify tags based on genre, theme and mood, and help provide extremely accurate recommendation and search results based on both lyrics and audio analysis; we’ve built an AI solution that detects and transcribes lyrics contained in a clip; and, to ensure an end-to-end solution, we’ve also built a dashboard that helps Songclip measure performance, learn from its users and adjust accordingly.
Our AI solutions got Songclip's capabilities from utilizing 4% of its database of clips to 100% in no time, which lets the company scale at a pace that would otherwise be impossible. After considering an in-house AI team, we’re proud to say Songclip will instead continue using STRV’s experts — who are there to help with anything, anytime.
“Songclip and STRV have been working together hand-in-hand for about 3.5 years, and we really do consider them a part of the team. They’re part of our regular, day to day work.”
- $23MTotal Raised in Funding
- 90%Approx. Accuracy of AI Model
- 10xIncrease in Clips Tagged
(Manual to AI)