🎧 This article is based on episode #126 of our podcast. You can listen to it here.
Max Engler is the co-founder of Dubly.AI, a young start-up from Hesse. Dubly AI is an AI dubbing tool that can translate videos into 28 languages - including lip-sync dubbing. The most exciting thing about the whole thing is that Dubly was built with low and no code. Max tells us how they did this and what they had to pay attention to when scaling the app.
The beginnings of Dubly AI
Max and his co-founder Simon started Dubly AI with the vision of lip-syncing videos and translating them with the original voice of the speakers. The idea was born out of the need to translate YouTube content into different languages without sacrificing the authenticity of the voice. With a clear focus on the use of AI and a deep understanding of the target group's needs, they created a platform that makes it possible to translate videos easily and efficiently.
No-code as the key to success
A central aspect of the development of Dubly AI was the use of no-code and low-code technologies. Max emphasizes that the combination of no-code tools and classic code forms the basis for the scalability and flexibility of their platform. The frontend and backend were initially developed with Bubble, while Xano is used for the database. This choice makes it possible to react quickly to market changes and continuously improve the platform.
Technology in use
Dubly AI uses a range of technologies to achieve the best possible translation results. For voice cloning and text-to-speech, Dubly AI relies on Eleven Labs, while Auphonic from Graz is used for audio enhancement. These carefully selected tools make it possible to maximize the quality and authenticity of the translations.
They also use a microservice architecture to solve tasks such as merging MP3 files from an S3 bucket or downloading YouTube videos. that cannot be implemented directly in Bubble. These processes are realized via AWS Lambda, a scalable tool that automatically starts new containers in response to API requests, executes the code and stops again. This allows Dubly to process thousands of requests simultaneously, which enables an efficient and flexible solution. The code required for these microservices was created using the AI assistant "Claude".
The future of Dubly AI
Despite the challenges associated with developing such a product, Dubly AI has shown that it is possible to achieve great things with a small team and limited resources. Max emphasizes the importance of focus and leveraging existing technologies to work quickly and efficiently. The platform is fully bootstrapped and has managed to hold its own against larger competitors with significantly more resources.
Conclusion
Dubly AI is an impressive example of how no-code and low-code technologies together with AI can lead to innovative and scalable solutions even faster. The platform not only offers an efficient way to translate video, but also shows how important it is to remain flexible and adaptable in order to succeed in a fast-moving market.
