An offline installation of Stable Diffusion is more than enough to cover any need or desire when it comes to generating images. However, that route demands some serious firepower from your computer. The alternative is to take a chance with one of the services that have emerged in recent months, and one of the most striking is Vana Portrait, which specializes in portrait generation. All it asks to get started is a pack of eight images, and a little patience…
Midjourney, Stable Diffusion, DALL-E… we've talked about those models to no end, but AI image generation finally exploded thanks to portrait apps like Lensa. The big advantage of these apps is that they lower the entry requirements: any user with a compatible smartphone or tablet can generate images in seconds.
But what about computers? The official demos of the main models haven't gone anywhere, however, one of the best equivalents to Lensa is Vana Portrait, which works directly in the browser.
Create Personalized AI Portraits with Vana Portrait
The first step to getting into Vana Portrait is creating an account. This takes just a few seconds and asks for nothing more than an email address. Then we must upload eight or more images of our face. Vana Portrait shares several examples of ideal images and which cases to avoid (face masks, more than one person in the image, out of frame, etc.). In essence, what Vana Portrait does is similar to training a mini-model with those images, to then apply different styles to them.
The total number of available styles amounts to 26, with examples like Pulp Manga, Cyberpunk, 80s Punk Rocker, Black and White, Cyborg, Flowers, and Home of the Flying Lizard, inspired by a certain TV series. Besides the styles section, we also find Gallery, where we can see all our generations, and Studio, which enables creating portraits using prompts.
All this sounds good on the surface, however, Vana Portrait has other details that should be mentioned. The first is that its privacy policy is not the best. The Privacy Policy section presents all levels of data collection, but if there is any problem with those terms or the algorithm generates a 'delicate' image, the user must contact via email, when the ideal would be to have automatic reporting tools.
The second is that Vana Portrait takes a long time to generate the initial portrait/model. Its official email mentions 24 hours, but some users have had to wait days to apply styles. Finally, Vana Portrait is not free. Its free demo is 50 credits, with a price of 5 dollars per additional pack of 50 credits.
Official site: Click here
