Qwant and Staan: A European Offensive Against Web Giants

Qwant and Its Evolution

Since its creation in 2013, Qwant has positioned itself as a European search engine concerned with privacy. Following its acquisition by Synfonium in 2023, the company has entered a new phase of development, focused on a more industrial and technological approach. The goal is no longer to compete directly with giants like Google, but to gradually build a credible and sustainable European alternative in the field of online search.

Qwant's strategy is based on three main pillars: offering a privacy-respecting search engine, strengthening its technological independence, and leveraging its expertise in search to meet the needs of businesses in artificial intelligence. This approach aims to create an ecosystem where personal data protection is at the forefront of concerns while developing innovative solutions for users and businesses.

European Search Perspective

In 2024, Qwant partnered with Ecosia to create European Search Perspective, a Franco-German joint venture aimed at developing European search technology and a web index. This initiative arose from the need to pool resources and expertise to build a solid alternative in a sector where investment and scale are crucial.

Developing search technology requires considerable resources to explore, index, and structure the web while continuously improving the quality of results. By joining forces, Qwant and Ecosia hope to build a common technology that will serve their consumer services as well as Staan, their API dedicated to professional uses. Lilo, another consumer service, will also benefit from this common technology, thereby strengthening the European offering against American giants.

Market State and Opportunities for Staan

The web access market is undergoing significant changes, largely due to the rise of generative AI, which requires reliable access to online content. However, this access is largely controlled by non-European companies, creating a dependency for European businesses developing AI services. Microsoft recently announced the discontinuation of its Bing Search API, further exacerbating this dependency.

Staan, Qwant's API, positions itself as a sustainable alternative to this dependency. It offers businesses a sovereign solution, allowing them to build their AI services without being subject to the unpredictable changes in access conditions imposed by dominant market players. This independence is crucial for companies looking to develop robust and sustainable AI solutions.

Staan: A European Gateway

Staan is a web search API designed for businesses, developers, and AI players who need to integrate web access into their services. Unlike a traditional search engine, it does not target end users directly.

This API allows AI applications to query the web for recent, structured, and sourced content, including relevant pages, snippets, metadata, and updated content. By providing direct access to reliable and up-to-date information, Staan enables developers to create more accurate and efficient AI applications.

Users and Use Cases of Staan

Since its launch, Staan has generated significant interest from companies of all sizes and sectors, including startups and players in artificial intelligence. Public organizations developing sovereign AI applications are also among its first clients.

The use cases for Staan are varied but share a common need for reliable access to recent and structured web content while respecting sovereignty. A notable example is Pappers IA, an assistant specializing in French legal and economic data, which uses Staan to ensure up-to-date and reliable information while protecting its clients from the implications of the Cloud Act. This ability to provide updated and secure data is essential for businesses that rely on accurate information for their operations.

Differentiation of Staan

While sovereignty is a major asset, Staan also stands out for the quality of the content it provides to AI applications. This content is recent, sourced, structured, and usable in modern architectures like RAG.

For developers, the challenge is to obtain not only search results but also usable information for AI systems, with sources, metadata, and snippets allowing for quick and relevant access to necessary information. This approach helps reduce development time and improve the accuracy of AI applications.

Quality Criteria in RAG Architectures

In a RAG architecture, the quality of the responses generated by AI heavily depends on the quality of the retrieved information. Essential criteria include the freshness of content, relevance of results, verifiability of sources, data structuring, and reliability of web access.

An AI that relies solely on its training data risks producing outdated or inaccurate responses. By accessing recent and identifiable information, the risk of hallucination is reduced, and the traceability of responses is improved. This ability to provide precise and verifiable information is crucial for the development of reliable and effective AI applications.

Prospects for the Web Search Market in Europe

In the next five years, the web search market for AI in Europe could undergo significant transformations. Qwant intends to play a central role in this evolution by consolidating its position as a credible European alternative and strengthening the technological autonomy of the continent in the face of the dominance of American players. This strategy aims to create a robust and independent European web search ecosystem capable of meeting the growing needs of businesses and developers in the field of AI.