We get this question occasionally from users of other Visual AI analysis tools, and the simple answer could be yes, it’s better. Nothing is as simple as black and white, so let us compare services from Goliaths like Google, IBM, Amazon and Microsoft with our David-like solution from Ximilar.

To say it simply, artificial intelligence vision got to a point, where it is easy not only to recognize objects in a photo, but also detect features of each thing. That creates a new universe of opportunities for real-world application in e-commerce & traditional industries alike. And Ximilar is a computer vision platform that digs deep into some pretty narrow use cases. So while the big solutions might be great in many ways, Ximilar might very well be the agile alternative.

Ximilar offers you a great cloud AI platform for training your custom image recognition models and advanced visual search services.

Ximilar is Not a Big Corporation

And that is a good thing. Because we keep things simple, streamlined, and we have time to listen to each customer’s needs. We also have the ability to implement new custom features in a timely manner. And we do it as fast as we can, widely benefiting both customers and us, freeing our manpower from manual work.

We at Ximilar create, and continuously improve, advanced visual search, image recognition services & image tools for businesses around the World. That happens in few areas:

• Ximilar recognition app – AI cloud platform for training custom image recognition models
• Ximilar annotate – image annotation tool for creating great datasets
• Visual Search – fashion, stock photos, real estate, home decor, cards, … search engines
• Ready-to-use services – fashion tagging, home-decor and collectibles recognition API
• Image tools – upscale your images and remove the background with API
• and in many cases the most challenging custom AI solutions

We are also not an enterprise that requires millions of users of its services to just stay afloat. See for example how many services were killed by Google. No. Rather than growth in quantity, our center of the universe is how precise we get, and how reliable & sustainable results we deliver. And how we can grow strong together with our customers, or we should rather say our partners.

Here is why Ximilar could be a solid alternative for you if you need to iterate quickly and reach reliable results in narrow fields. Or if you simply need someone who takes your idea further and finds an AI solution to deliver value to your business.

1 – We are focused AI Team

We craft our features to perfection, and we test & use them ourselves. We continuously improve our application for everybody to benefit from new findings in AI vision industry. And we also do things that customers ask for, we don’t just sell access to a platform.

2 – We are an independent company

These days, many companies are created to be acquired. They are created to grow no matter the sustainability of such growth. We are different. Our customers like that we would not disappear tomorrow — getting acquired by a giant and then dissolved into some unreachable feature of some huge app suite is not our target.

3 – We innovate faster

We don’t have a large team and therefore decisions are quick. We are a team of remote professionals working in a field that we truly love and would like to explore to the edge of possibilities. It’s a lot of fun to work on our customers’ challenging tasks. And we are happy to customize any feature. The customer’s budget is the only limit.

4 – Save expenses on AI

Our AI solutions are significantly cheaper than the solutions of big AI players. We are able to save you a lot of money on training and deploying your custom models. For example, training and deploying a model on Google Vertex AI can cost you thousands of dollars, without even calling the API. For Vertex AI AutoML models you are paying for training, deploying and calling a model. Similar pricing for features can be applied to Amazon Rekognition and Azure Custom Vision services. With Amazon Rekognition you are also paying for each hour your model is deployed! On the other hand, AI models built via our platform are trained and deployed for free! You are paying just for calling the API. No more hidden costs.

Head-to-Head Comparison

Narrow Field vs. Generic AI

This one is personal. You would see a lot of simple AI applications, like detecting a cat and a dog in a given — well lit & well shot — picture. But in reality, the bread and butter of applied visual AI is narrow field recognition and analysis of large volumes of images, where the customer needs pretty high accuracy on a specific subject. For example, detect a type of screw on a blurry cellphone photo, shot in bad lighting conditions.

Unlike the giants who mostly sell you ready-made solutions that you can hardly bend to meet your needs, Ximilar is in the other end of the spectrum, brainstorming with customers about how to solve the use case that they have. Being their partner in the path to success.

Examples of such narrow use cases are

• Detecting coffee grounds in a cup – for a customer who receives millions of images to their mobile app used to foretell the future for its users. You wouldn’t believe how many users in coffee-drinking countries use such an app.

• Recognition of trading cards from photo – A cool use-case that was a dream of every geek. Not anymore. Simply snap a photo of a sports card or a game like Pokémon, and the app will identify a card and return a price listed on eBay. You can build your own portfolio tracker and much more with Ximilar.
• Give me a quality rating of a photo – this one was brought up by a hotel reservation site and real estate company. They need to detect the best photos of a property, while the photos are often delivered by a re-seller, or a hotel owner and might not be well shot. And we all know that good photos sell better. Ximilar can help even there with upscaling images and improving their quality.

Lower Price for Higher Accuracy

While the examples above might be fun to read, let’s get to real facts, hardcore numbers and actual user feedback. Because that is a requirement for any business to base its thoughts on. Here are some real-life examples of our customer experiences.

• Ximilar Recognition is cheaper and has comparable accuracy as Microsoft Custom Vision, Amazon Rekognition, Google Vertex AI and IBM Watson. At least several of our customers, and users of the Ximilar App, achieve even better accuracy than with the big cloud solutions. Ximilar allows users to control various parameters of training from a simple GUI.

• UX of Ximilar App is extremely easy to use, also reported by our customers, saying: “Ximilar has a shallow learning curve in comparison to others”. Connection to the API and integration to your systems and apps is easy.
• Ximilar has advanced features for tuning of your recognition tasks which no other services provide — flips, rotations, etc.

• Ximilar Product Similarity and Custom Similarity are unique services for finding visually similar alternatives in fashion, home decor and other image collections
• Ximilar is much more flexible as we are willing to improve our service for your needs – e.g. add more tags to our models — according to your requirements and keep it attached to your data exclusively
• We are cheaper — Google AutoML Vision/Vertex AI is significantly much more expensive than our solution
• Ximilar Fashion Tagging is at the top of abilities in fashion object recognition
• Elaborate management of tags & categories for more projects of higher complexity — we are the only system we know of, that enables users to share training data between categorisation and tagging tasks, chaining recognition models into one API…
• Ximilar, unlike the big competition, is able to install the system on-premise, giving you better control over the system, do a lot of flexible customizations

This is just a brief summary of what we see as benefits for you if you use Ximilar as your partner for pioneering the AI world. We see it now as really just the beginning of all the possibilities that might come in the future of automation and machine learning abilities. We have been around for many years now and Ximilar would surely be around for the years to come. Backing you on the way. Enjoying the exploration.