Voice Agent Compliance & QA
Evaluate calls handled by AI voice agents against a policy you write. Every finding shows the transcript quote it is based on, so a reviewer can confirm or dismiss it in seconds.
Free during early access, up to 200 calls per month.
We guarantee a full refund.
Hello, this call is recorded for quality purposes.
How results are worded
GoldNovaAI produces QA signals for a human reviewer. It does not certify that a call is legally compliant.
No enabled rule found a problem. This is not a guarantee that the call was fine.
Something needs a person: an uncertain match, or a check that depends on who spoke when the speaker is unknown.
A rule matched, with at least one verbatim quote from the transcript as evidence.
Send a call, read the result
Ingestion is idempotent on your call id, so retries never create duplicates. Evaluation runs right after the call is accepted.
curl -X POST https://goldnovaai.com/api/v1/calls \
-H "Authorization: Bearer $GOLDNOVA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"externalCallId": "call-1042",
"agent": { "id": "support-bot-1", "name": "Support bot" },
"transcript": [
{ "speaker": "agent", "text": "Hello, this call is recorded for quality purposes.", "startTime": 0, "endTime": 4 },
{ "speaker": "customer", "text": "Okay.", "startTime": 4, "endTime": 5 },
{ "speaker": "agent", "text": "We guarantee a full refund.", "startTime": 41.2, "endTime": 44.0 }
]
}'{
"status": "evaluated",
"evaluation": { "result": "violation", "resultLabel": "Potential violation", "qaScore": 60 },
"findings": [{
"rule": "Prohibited claims",
"severity": "high",
"status": "violation",
"evidence": [{ "quote": "We guarantee a full refund.", "startTime": 41.2, "endTime": 44.0 }]
}],
"notice": "QA signal for human review. Findings are automated indications, not legal or compliance determinations."
}The request returns the call id and status; the result is read with GET /api/v1/calls/{id}. The example shows one finding: with the starter policy the same call would also get review recommendations for the script and closing-wording rules.
Limits: 1.5 MB per request, 5,000 transcript segments, 120 requests per minute per API key. Keep keys on your server.
Availability
| Capability | Status |
|---|---|
| Generic API for transcripts, dashboard, call reports, policy editor | Available |
| Recording upload with speech-to-text (speakers not labelled) | Available |
| Optional AI assessment for script adherence and failed promised actions | Available when enabled for the deployment |
| Vapi, Retell and Twilio webhooks | Developer preview not yet tested with live provider accounts |
What it is not
GoldNovaAI is not legal advice and does not certify regulatory compliance. You remain responsible for call-recording consent and for the regulations that apply to your business. We do not currently hold SOC 2, ISO 27001 or similar certifications.