Users are constantly being bombarded with ever-growing volumes of content, such as larger product offerings from retailers or exponentially growing numbers videos to sift and watch. As users have more choice about where to spend their time on different apps and sites, and where to spend or invest their money, it is essential for companies to be able to effectively match them to relevant content quickly and reliably.
High-quality recommendations aren't just a bonus feature on apps and websites, they are a critical function. Most companies, however, are not recommendation experts and don't have or want the internal expertise to build and maintain complex, modern recommendation systems. Luckily, reinventing this technology isn't necessary!
Lucid provides APIs, software, and support to help you build fast and accurate recommendation and search tools for your site and apps.
Lucid specializes in contextual understanding. Our technology and APIs focus on combining multiple data sources together to identify key moments, trends, and objects in your content. We're able to build this understanding by structuring your data in a highly connected graph format, which we can then train graph neural networks on. With your data in network form, our AI then is able to pick out relationships, anomalies, trends, and patterns that traditional methods may have missed.
Building graphs, training graph neural networks, and analyzing these networks with modern algorithms is hard work. Lucid simplifies this process and takes care of the complexity. As a user of the APIs, you just need to put your data into a simple format and our backend will handle the rest.
You can connect our recommendation and trends tech to many different domains:
- News and media
- Enterprise data
- Boards, postings, and user content
Lucid's APIs are centered around graph representations of your data, which allow us to train graph neural networks. In a graph format, all of your data is represented as either a vertex or edge. For example, in an e-commerce application, each products and user will have their own vertex, and edges represent the interactions between them, such as purchases, views, or shopping cart actions.
One your data is in the graph representation, a graph neural network will be trained. Lucid handles selecting the optimal parameters for training these neural networks, and will continuously optimize each network as more data becomes available.
While graph networks are easy to understand for simple cases, they can also be applied to a surprising number different domains. For example, Lucid is able to decompose videos into a graph representation to do complex AI analysis on them. Scenes, segments, people, and objects can all be represented as a vertex, with edges indicating which combinations appear together. In addition, videos content can be connected with our larger knowledge graph of data collected from our video corpus. This lets us ask questions such as:
- Do we know who the actors / characters are in this scene?
- If we don't know their names, have we seen them before in other videos?
- Are the topics in a video scene related to trends in the news or social media?
- How does this scene fit into the larger context of the video?
- What are the key moments a particular user may want to see in this video?
Whether you're using our tech with e-commerce data, news and media, or video, a key requirement is being able to learn over time. Lucid's graph is designed to monitor user interactions to improve its accuracy in real-time. For example, user and product trends can change very quickly, and our graph algorithms aim to optimize the most up-to-date recommendations and patterns.
Lucid connects your users with what they want to see and interact with. Whether you're serving e-commerce products, news and media, video, or other content, we can analyze and organize it to home in on what each user is most interested in.
Our tech will let you better serve your users:
- Provide user-optimized personal recommendations
- Connect them with local trends
- Increase discoverability by serving content outside what they may search on their own
- Give them powerful contextual search capabilities
Our tools are all about making your business more agile and more profitable. We take care of the backend details for recommendations, trends, and contextual understanding, so your developers and marketing teams can focus elsewhere.
With our API, your business can:
- Get personalized user recommendations in real-time
- Track and promote trending items
- Get multiple recommendations for easy A/B testing
- Extract insights on user behavior
- Extract insights on your product lines and services
Lucid's software tools, portal, and APIs make it easy for developers to get set up and running in minutes. Integrating with your site and apps is fast and seemless:
- Simple API endpoints: we've abstracted out to complexity into an easy API
- Easy framework integration: we mesh with many popular backend and frontend libraries, integrations, and frameworks
- Secure: we protect your data and give you control over what we see
- State-of-the art: our algorithms use the most modern graph network tech, with continuous updates to improve performance
- Highly optimized: low-latency results and low-hassle for integration
- Detailed documentation: API references, how-to guides, and case studies
- Enterprise support: we're always available to support your development and answer questions