Analyze the calls - FormBharo

Agent analytics

GET /api/v1/agents/{agent_id}/analytics returns stats for one agent.

curl

curl -G "https://api.formbharo.artpark.ai/api/v1/agents/AGENT_ID/analytics" \
  -H "Authorization: Bearer <api-key>" \
  --data-urlencode "since=1700000000" \
  --data-urlencode "until=1700086400" \
  --data-urlencode "status=complete" \
  --data-urlencode "source=web"

Query parameter	Description
`status`	`in_progress`, `complete`, `screened_out`, or `failed`
`source`	The source that started the call
`since`	Unix timestamp: `created_at >= since`
`until`	Unix timestamp: `created_at <= until`

Workspace analytics

GET /api/v1/workspaces/{workspace_id}/analytics returns the same stats across all agents in a workspace, plus a per-agent breakdown. Workspace admins only.

curl

curl -G "https://api.formbharo.artpark.ai/api/v1/workspaces/WORKSPACE_ID/analytics" \
  -H "Authorization: Bearer <api-key>" \
  --data-urlencode "since=1700000000"

Reading the numbers

Stat	Meaning
`attempted`	Calls started, in any status
`completed`	Calls that reached a defined ending, including screened out ones
`screened_out`	Calls ended early because an answer matched `end_call_values`
`completion_rate`	`completed / attempted`, between `0` and `1`. `0` when `attempted` is zero

To count calls that answered every question, subtract screened_out from completed. The response also carries average call duration, voice latency, and cost per call and per turn. See the API reference for those fields. Every average is null when the window has no usable calls.