Vapi call analytics

Chart quality and latency across your Vapi calls by minute, hour or day, with every point opening onto the conversations underneath it.

Vapi reports what your calls cost and how they ended. This reports whether they went well, on the same calls, and lets you chart it beside your own numbers.

What you can pull

Quality	Scores by rule and by area, bucketed by minute, hour or day.
Latency	Reply times measured from Vapi's millisecond timestamps.
Abandonment	Where callers gave up, grouped by the reason they gave up.
Coverage	How many calls were scored, and by which rule version.

All of it is available as JSON, so it goes into whatever dashboard you already run instead of becoming a tab nobody opens.

Why the buckets matter

A daily average hides a bad hour. An agent that fails ten percent of calls between six and seven, and none the rest of the day, reads as a good agent at day resolution.

Minute buckets make a deploy visible. The number moves when the release lands, not when somebody files a ticket about it a week later.

Every point opens

A chart you cannot interrogate is a chart people argue with. Each reading here leads back to the calls that produced it, and each call to the turn a rule fired on.

Scores are frozen under the rule version that produced them, so a chart of last quarter does not change when you edit a rule today.