TLDR: Qdrant has launched Cloud Inference, a fully managed service designed to unify embedding generation and vector search for text and images. This innovation aims to significantly boost speed, reduce costs, and simplify the development of AI applications by offering a single API call for multimodal data processing.
Qdrant, a prominent provider of high-performance, open-source vector search solutions, has officially announced the launch of Qdrant Cloud Inference. This new fully managed service is set to revolutionize how developers handle multimodal data by enabling simultaneous search across both text and image vectors. The core objective of Cloud Inference is to deliver higher speed and lower operational costs for AI-driven applications.
Traditionally, developers building artificial intelligence workflows have faced the challenge of managing separate pipelines for embedding unstructured documents and images. This often leads to fragmented processes where text and visual data, even when part of the same document, must be separated and processed distinctly. Qdrant Cloud Inference addresses this inefficiency head-on by integrating embedding generation and vector search into a single, cohesive tool.
André Zayarni, CEO and co-founder of Qdrant, highlighted the transformative nature of this integration, stating, “Traditionally, embedding generation and vector search have been handled separately in developer workflows. With Qdrant Cloud Inference, it feels like a single tool: one [application programming interface] call with optimal resources for each component.” This unified approach allows developers to combine dense and sparse unstructured documents alongside images within one environment, leading to reduced latency, lower network costs, and a simplified search experience.
Vector databases, such as Qdrant, are fundamental to modern AI, serving as essential ‘memory’ for AI assistants and agents. They enable these intelligent systems to retain context over extended periods by acting as a semantic memory, helping AI to ‘remember’ and retrieve information based on meaning rather than just keywords. This capability is crucial for advanced AI applications like Retrieval-Augmented Generation (RAG) and multi-agent systems.
Also Read:
- Clarifai Advances Agentic AI with New Server Hosting and OpenAI Integration
- NVIDIA DGX Cloud on AWS and Amazon Bedrock Custom Model Import Streamline Generative AI Development
At launch, Qdrant Cloud Inference supports six curated models for multimodal and hybrid search applications. The service offers a generous free tier, providing users with up to 5 million free tokens per model per month, along with unlimited BM25 tokens. Inference capabilities are automatically enabled for new paid clusters running version 1.14.0 or higher, and existing clusters can activate the feature with a simple click in the Qdrant Cloud console. Qdrant is also hosting an inferencing webinar on July 31 to provide further insights and demonstrations of the new service.
