Qdrant Cloud Inference provides native embedding generation and storage capabilities within managed Qdrant Cloud clusters, enabling users to perform multimodal and hybrid search without relying on external pipelines or model servers. This service is designed for applications that require low-latency, in-region embedding and retrieval, such as real-time search or recommendation systems, and supports both text and image data. By generating embeddings directly inside the network of the Qdrant Cloud cluster, it eliminates the need for external data transfer and reduces latency, making it suitable for scenarios where delays or egress overhead are critical concerns.
The platform supports a range of model types, including dense models like all-MiniLM-L6-v2 for semantic matching, sparse models such as splade-pp-en-v1 or bm25 for keyword recall, and CLIP-style models for handling both image and text data. This flexibility allows users to implement vector search tailored to their specific needs, whether for hybrid or multimodal use cases. Embeddings can be generated for both text and image inputs, and the system is compatible with AWS, Azure, and GCP cloud providers in the US region.
Qdrant Cloud Inference is accessible through a single API, streamlining the embedding and search workflow within the managed cluster. For new clusters, the feature is enabled by default, while users with older clusters can activate it via the Qdrant Cloud Console. The platform also allows for the use of external model providers, in which case embeddings are generated outside the cluster network.
Pricing for Qdrant Cloud Inference is based on a per-token billing model, with costs varying depending on the selected model. Paid Qdrant Cloud users receive up to 5 million free tokens per month for each model, and several models are available entirely free of charge with no token limits. Free models and external providers can be used in free Qdrant Cloud clusters, while paid models require a paid cluster. The service is positioned as a solution for advanced search, recommendation systems, retrieval-augmented generation, data analysis, anomaly detection, and AI agent applications across industries such as e-commerce, legal tech, hospitality, HR, and healthcare.
Qdrant Cloud Inference sits in PulseGate's AI & ML category. It focuses on generating and storing embeddings for advanced search and retrieval in vector databases. Qdrant Cloud Inference is a B2B product aimed at developers, data scientists, AI engineers. The product follows a commercial open-source model under the Open Source license. It runs on the web and API.
Qdrant builds and maintains Qdrant Cloud Inference, and the product first shipped in 2024. Among its 7 catalogued features are native embeddings, vector search, and multimodal support. It exposes integrations via a public API.
Latest indexed changes and source events
Qdrant Cloud Inference verified by the PulseGate indexer
Other apps tracked under the same category.
