Spam calls are filtered automatically. The platform uses three signals layered on top of each other — carrier reputation, live agent classification, and post-call analysis. Spam calls don't count toward your billed call quota.
How spam filtering works
| Layer | When | What it does |
|---|---|---|
| Carrier reputation | Before the AI answers | Every inbound call carries a spam score from the carrier (Telnyx). If the score is above the threshold (default 0.7), the call is routed to spam handling without the AI ever picking up. No bot voice, no minutes used. |
| Live classification | During the call | If the call gets past the carrier check but the AI hears clearly-bot behavior — repeating prompts, no human turn-taking, classic spam patterns — it can mark the call as spam mid-conversation using a built-in tool. The call ends and is tagged. |
| Post-call analysis | After the call | When a call ends, an analysis pass classifies the intent. Calls whose intent is determined to be spam are flagged after the fact. |
What this means for your bill
Spam calls — caught by any of the three layers — are not billed. Only legitimate, billable calls count toward your monthly quota and overage rate.
What you can see
In the Phone inbox, spam-flagged calls appear with a Spam badge. You can review the transcript (when there was one) to verify, or filter them out of your default view.
What's NOT supported today
- Manual blocklist — there's no UI to add a specific number to a do-not-answer list. The platform relies on the auto-filter signals above.
- Custom spam threshold — the 0.7 carrier threshold is fixed; you can't tighten or loosen it per account.
- Allowlist — there's no "always answer this number" override.
Marking a call as spam after the fact
If a call slipped through and was clearly spam:
Open the call
From the Phone inbox.
Mark as spam
Tap Mark as spam.
The classification feeds back into your account's analytics; today it doesn't yet block the number from calling again.