nSearch
AI Search and Intelligence Platform.
AI search, recommendations, agentic AI, RAG, anomaly detection, fraud detection, and vision AI on one vector intelligence platform. Deploy in the cloud, on-premises, or at the edge.
50+
Languages
Multilingual semantic search
<5ms
Query Latency
At billion-vector scale
1B+
Objects
Indexed on a single node
3
Deployment Tiers
Cloud, Enterprise, and Edge
Core Capabilities
What's under the hood.
Vector Search
High-recall ANN search across dense embeddings at any scale.
Text Search
Sparse BM25 keyword search with language-aware tokenization.
Multi-vector Search
Index and query across multiple embedding spaces per document.
Context Graph
Graph-aware retrieval that follows semantic relationships.
Progressive Filtering
Metadata filtering runs inside the ANN graph. No post-retrieval pass.
Quantisation
Proprietary compression that preserves accuracy while reducing storage.
Intelligent Search for Every Workload
nSearch powers the full spectrum of AI intelligence use cases from a single vector engine.
AI Search
Search by meaning, not keywords. Handle synonyms, multi-lingual queries, and complex intent with dense vector retrieval and metadata filtering at any scale.
Recommendations
Real-time personalized recommendations using item and user embeddings. Content-based, collaborative, and hybrid, with cold-start support from day one.
Agentic AI and RAG
Give agents persistent long-term memory across sessions. High-recall retrieval means agents reason on facts, not approximations, completing workflows more reliably.
Anomaly Detection
Detect anomalies in transactions, sensor data, images, and logs using vector similarity. No labeled training data required to get started.
Fraud Detection
Real-time transaction screening and behavioral pattern matching. Flag suspicious sequences before they complete, at the throughput of production payment systems.
Vision AI
Image similarity search, multimodal retrieval, and video analysis at billion-scale. Works with CLIP, ImageBind, and any custom vision encoder.
Start free. Scale when you're ready.
Start free and scale as your AI applications grow.
*Overage Billing: Overage charges apply per full block beyond your quota. Pro: $99 per 1M extra queries. Scale: $999 per 10M extra queries. Starter has no overage and queries are capped.
*Upgrade Proration: Upgrading Pro to Scale within the same billing cycle? You only pay the $900 difference. Upgrades from Starter are charged at full price.
*Hybrid Search & Object Counting: Each hybrid record is stored as two objects internally (one dense and one sparse). This halves your effective capacity. For example, on Starter (1M objects), you can store up to 500K hybrid records.
Deploy on your infrastructure.
Monthly billing. No setup fees. Scale up or down at any time.
Plans at a Glance
| Plan | Monthly | Architecture | Capacity |
|---|---|---|---|
| Single Node | $2,000 | Single node deployment | Up to 100M items |
| High Availability | $5,000 | Multi-AZ, cross-zone read replica | Scales with nodes |
| Cluster | $20,000 | Full distributed cluster | Unlimited vectors |
Vector search on your hardware
Run a full Endee vector database on ARM, x86, and Android edge hardware. Purpose-built for air-gapped environments, sovereign deployments, and devices that never touch the cloud. Licensing is tailored to device count, fleet size, and update cadence.
Independently Verified Performance
Outperforms all other vector databases across every metric
Independently verified. Higher throughput, lower latency, and better recall than every other tested vector database.
Frequently Asked Questions
Common questions about nSearch and the Endee AI platform.
Start building with nSearch
Free tier includes 1M vectors and 10K queries per day. No credit card required. Scale to billions as your applications grow.