Another AI-native application has gone viral. Recently, the AI image enhancement app Magnific AI has garnered widespread attention, attracting 400,000 registered users just over a month after its launch.
Magnific AI not only uses generative AI technology to upscale images, but it also enhances the resolution with just one click, making the original image clearer and more textured. Even Elon Musk has praised this app on social media.
In this era of large language models, new AI creation tools are gradually emerging. Many users are now combining tools like Midjourney, Magnific AI, and Runway to aid in video creation with AI.
Magnific AI hails from Spain and was created by a team of just two founders: CEO Javi López and CTO Emilio Nicolás. Currently, most of the app’s users are international, with only 5% coming from Spain.
Magnific AI was born out of an unexpected hobby for López and Nicolás. After OpenAI released DALL-E 2 in 2022, López became interested in generative AI and built a community around it in his spare time. By October 2023, López and Nicolás decided to fully commit to their idea and spent a month developing the first generation of Magnific AI.
López revealed that Magnific AI uses an open stable diffusion model. They also discovered that tweaking parameters and source code led to some surprising new product features. Like many AI tools, Magnific AI requires users to input natural language prompts to describe the desired image effects.
Magnific AI can also upgrade emojis into 3D portraits. Unlike other AIGC tools like Pika and Midjourney, Magnific AI's prompts can be broad descriptions or specific instructions regarding the desired image movement, such as enlarging or shrinking a particular area.
Additionally, the Magnific AI tool interface includes options to adjust values like "Creativity," "HDR," and "Resemblance" (similarity to the reference image), as well as fractality (movement direction) to ensure image clarity and content consistency.
Creativity affects the AI's "hallucinations," influencing the details of the generated image. HDR refers to the image’s clarity and level of detail, while Resemblance determines how closely the generated image matches the reference. Increasing this value makes the AI-generated image more similar to the original.
One user used the prompt "robot portrait, futuristic, grey, black, yellow" to magnify a tiny camera on a robot's head into a detailed device with lens details.
mike82839