MassEngine
Live 20,000,000-Vector Multi-Cluster Studio

Same 20M Records. 3 Hardware Setups Live.

Query the exact same 20,000,000 English Wikipedia passages (1024-d embeddings, 83.46 GiB logical corpus) across three live cluster configurations simultaneously: 2-Node High-End In-Memory, 2-Node High-End NVMe, and 4-Node Low-End NVMe.

Target Cluster Setup:
Setup 1 • In-Memory
2-Node High-End In-Memory
2 Nodes • 8 Shards (RF=2) • Full RAM Resident (~39 GiB/node)
— ms
100% Top-K Match
Waiting for query...
Setup 2 • Direct NVMe
2-Node High-End NVMe
2 Nodes • 8 Shards (RF=2) • Memory-Capped NVMe Tiering
— ms
100% Top-K Match
Waiting for query...
Setup 3 • Commodity NVMe
4-Node Low-End NVMe
4 Nodes • 8 Shards (RF=1) • 1 vCPU / 6.8 GiB RAM per Node
— ms
100% Top-K Match
Waiting for query...
One-Click Preset Queries:
Corpus: 20,000,000 × 1024-d Vectors (8 Shards) • Mode: All 3 Setups Side-by-Side
Status: Ready
Live Ranked Hits (0)
Generated API Payload (POST /v1/search) 
{
  "vector": [
    0.0241,
    -0.0188,
    0.0412,
    -0.0093,
    "... (1024 float32 dims)"
  ],
  "k": 8,
  "nprobe": 16,
  "rerank_pool": 64,
  "metadata_filter": {
    "category": { "$eq": "science" }
  }
}
Real-Time Observability & Metering

Customer Usage & Telemetry Portal

Every workspace includes built-in latency percentile histograms, partition-level byte metering, and real-time usage tracking (GET /v1/tenants/{tenant}/telemetry). Select any of the 3 demo clusters below to inspect its live telemetry stream:

Workspace Telemetry • demo-cloud
• LIVE TELEMETRY STREAM
Live Vectors (Logical)
20,000,000
Quota: 50,000,000 max
Logical Storage Size
83.46 GiB
8 Shards • Logical bytes metered
Search Latency (P50 / P95)
2.0 / 8.8 ms
Avg: 2.34 ms • P99: 9.8 ms
Metered Read / Write Units
566 / 20.0M
Rate limit: 500 QPS (0 throttled)

Search Latency Distribution (ms) Live Histogram

≤0.5
≤1
≤2.5
≤5
≤10
≤25
≤50
≤100

Real-Time Usage Counters Per-Workspace Metering

Logical Vector + Metadata Footprint 83.46 GiB
Cumulative Search Read Units 566
Cumulative Ingest Write Units 20,000,000
Hard Spend & Quota Guardrails Active (0 Throttled)
Active Tenant Partitions & Routing Mode (GET /v1/partitions)
Scope / Partition Routing Mode Live Vectors Logical Size Status
demo-cloud / default supercell (8 shards) 20,000,000 83.46 GiB Healthy