Analytics & dashboards
You can't improve an agent you can't measure
Without hard numbers, nobody knows what share of calls the agent actually contains, why the rest escalate, how callers feel, or what a conversation truly costs. Every decision is a guess, and margin quietly erodes one untracked minute at a time.
What's in the box.Each one in the product today.
- 01
Containment & escalation
See what share of conversations the agent resolves end-to-end and why the rest hand off.
- 02
Sentiment & duration
Track how callers feel and how long turns and calls run, per agent and over time.
- 03
Itemized cost & margin
Break every conversation into STT, LLM, TTS, and transport COGS — and the margin left over.
- 04
Custom dashboards
Compose metric, time-series, and breakdown widgets into a saved dashboard.
- 05
Per-agent rendering
Render the same dashboard per agent or across the workspace — build once, view anywhere.
Every conversation broken into STT, LLM, TTS, and transport COGS
Mirakash prices each stage of a conversation — transcription, model, speech, and transport — into a per-call COGS and the margin left against what you charge. So unit economics are a number on a dashboard, not a spreadsheet you rebuild each month.
- Call
- C-2204
- Agent
- Admissions
- STT—
- LLM—
- TTS—
- Transport—
- starting…
One signal stream feeds every metric you define
Containment, escalation, sentiment, duration, and itemized cost land on one stream per conversation. From it you compose metric, time-series, and breakdown widgets onto a saved dashboard, then render it per agent or across the workspace and window it however you need.
- Stream
- one per call
- Window
- day
- Escalation · metric
- Duration · time series
- Sentiment · by agent
- Cost · by stage
- starting…
Explore the platform.
It works with the rest of Mirakash. Here is what connects to it.
- QA, scorecards & human handoffLLM-judge scorecards, flagging rules, a review queue, and human-in-the-loop escalation that resumes the caller.
- API & connectorsServer actions and an API, real env-gated connectors, {{VAR}} interpolation, and agent tools that call your endpoints.
- Realtime voice pipelineSub-second STT→LLM→TTS with barge-in, on ~11 pluggable provider adapters or native Gemini Live.
- Use casesA real conversation for six desks, start to finish.
- SecurityRoles, redaction, retention and the audit trail.
- PricingPer minute or per resolution, and what each includes.