Linkup, a French startup, is building a solution to provide developers with access to trusted web content, helping enhance large language models (LLMs) with reliable, up-to-date information. This approach addresses a significant issue in AI development known as Retrieval-Augmented Generation (RAG), where external data is used to improve the quality and accuracy of AI-generated answers.
While many AI companies use web scraping to collect data for training models, this practice is increasingly under scrutiny, especially when there’s no financial agreement with the content publishers. Legal cases, such as the lawsuit between OpenAI and The New York Times, highlight the concerns over the use of content without compensation, prompting many companies to sign licensing deals with publishers.
Linkup was founded at a time when AI companies, like OpenAI, started establishing agreements with publishers to pay for access to their content. Linkup aims to act as a bridge between content providers and AI developers by facilitating legal content licensing. According to Philippe Mizrahi, Linkup’s CEO, the goal is to create a mutually beneficial ecosystem for both AI developers and content creators.
For content publishers, the options for handling AI's need for data are limited. They can block web scrapers using a robots.txt file, take legal action against companies violating copyright, or choose to license their content. However, licensing content to AI developers can be complicated, especially for smaller publishers who may not have the resources to pursue lawsuits or engage in large-scale licensing agreements.
Linkup’s solution is to serve as a marketplace where content publishers can license their data to AI companies. The company signs licensing deals with publishers and integrates with their content management systems (CMS), allowing it to access content without scraping. Publishers are then paid based on how frequently their content is used by Linkup's clients.
Mizrahi explains that Linkup focuses on AI applications in business, where developers need to enhance their models with external information. For example, one of Linkup’s customers created an AI application for sales teams. By integrating Linkup’s services, they were able to provide their sales reps with real-time, high-quality data to generate personalized sales pitches.
In addition to news websites, Linkup also works with knowledge databases, such as Statista and Xerfi, to provide more specialized content for AI models. While Linkup competes with other startups like ScalePost, which also focuses on licensing deals with publishers, it stands out by offering a more streamlined solution for accessing premium content legally.
Having raised €3 million in seed funding, Linkup plans to expand its team and continue growing its platform, helping AI developers enhance their models with legally sourced, high-quality content.
