sarvamvoice
Performance Monitor

Live Service Metrics

Real-time tracking of transaction statistics, pipeline latencies, and speech synthesis performance.

Last updated: just now
Total Queries

0

Total API calls processed

Avg Latency

End-to-end pipeline time

Success Rate

100.0%

Target SLA: > 95%

Pass / Fail

0

Passed

0

Failed

Token Cost (USD)

$0.0000

Avg: $0.000000

Pipeline Latency Breakdown
Max: 5.0s
Speech-To-TextSaaras v3
0.000s
FAQ RetrievalTF-IDF Similarity
0.000s
LLM InferenceSarvam-30B
0.000s
Text-To-SpeechBulbul v3
0.000s
Total avg pipeline time
SLA Status
  • Response Speed

    Avg latency under 5 seconds

  • System Accuracy

    Query success rate > 95%

  • RAG Relevancy

    FAQ vector index healthy

  • TTS Audio Quality

    Bulbul v3 synthesis active

System Operational

Quick Navigation

Voice Agent
Landing Home
Agent Evaluation Suites

No evaluation results loaded yet.

Agent Execution Trace Viewer

No traces fetched yet.

Enter a Session ID above or perform an interaction in the playground to view execution flows.

Voice Pipeline Architecture
STEP 01

Speech-to-Text

Saaras v3

Indic ASR with dialect support

STEP 02

RAG Retrieval

TF-IDF / SQLite

Sub-50ms FAQ vector lookup

STEP 03

LLM Inference

Sarvam-30B

Sovereign Indic chat model

STEP 04

Text-to-Speech

Bulbul v3

Natural voice synthesis