#153
A+
Rank #3
Qdrant
Qdrant is a high-performance vector similarity search database built in Rust for AI/ML applications like RAG and recommendation systems. Its standout trait is being cloud-native from the ground up, with best-in-class container design and clustering.
89.3
/ 100
⚙ Full Stack
A single Rust binary exposing gRPC and REST APIs, with an HNSW-based vector index, on-disk payload storage, and native Raft-based clustering/sharding for distributed deployments.
📈 Scaling Analysis
Qdrant scales near-linearly by adding nodes and letting its Raft-based consensus handle sharding and replication, with the same binary running identically from a laptop to a multi-node production cluster — a design that makes it one of the most genuinely cloud-native apps in this batch.
🚀 Running on Nexlayer
Deploy as a single pod for small workloads or a StatefulSet-style pod group for clustering, using <podName>.pod:6335 for the internal Raft/cluster gossip port; the official image is minimal and health-check friendly, and mirrors cleanly through mirror.gcr.io. Persistent volumes are needed per node for the on-disk index.