🤖

RAG Pipeline Analytics & Knowledge Base Monitor

Real-time observability into embedding spaces, retrieval quality, and LLM generation pipelines

Total Documents Indexed
12,847
↑ 2.3% from last week
Total Queries Processed
156,234
↑ 18.5% from last month
Avg Retrieval Latency
45ms
↓ 12% improvement
Avg Relevance Score
0.78
↑ 5 points from baseline
Hallucination Rate
3.2%
↑ 0.1% from last month
Answer Accuracy
92.1%
↑ 1.8% improvement
Chunk Hit Rate
87.3%
↑ 3.2% from baseline
Embedding Dimensions
1536
OpenAI Ada-3
Query Volume & Relevance Score Timeline (3 Months)
Cosine Similarity Matrix: Document Category Overlap
2D Embedding Space (t-SNE Projection) with Retrieved Chunks
Document Distribution by Category
Average Relevance by Category
Query Category × Document Category Relevance Heatmap
Top-K Retrieved Chunks & Relevance Distribution
Precision@K
0.85
Recall@K
0.72
MRR
0.81
NDCG@10
0.88
RAG Pipeline Funnel: Volume & Drop-off
Latency Breakdown by Pipeline Stage
Stage-wise Metrics
Stage Volume Latency (ms) Success Rate
Relevance vs Answer Quality
Faithfulness Score Distribution
Answer Quality by Document Category
Token Usage Over Time (Input vs Output)
Cost Tracking (Monthly)
Avg Tokens per Query Trend
MCP Tools Connected
24
Healthy
Resources Available
156
Healthy
Prompts Cached
412
87% Hit Rate
Avg Tool Latency
32ms
Optimal
MCP Tool Invocation Timeline
MCP Tool Usage Frequency
Tool Latency Distribution
Recent Queries & Performance Metrics
Query Text Category Top Doc Retrieved Relevance Score Latency (ms) Tokens Used Quality Rating