Hi Retell team, We’ve recently experienced a spam attack where calls were triggered from many different phone numbers and connected to our AI Voice Agent. In multiple calls, the exact same prerecorded audio was played and the agent continued the conversation, which obviously generates costs. This is not the first time this has happened. We’ll attach a sample recording for reference. Questions: Is there any way to block or detect this type of behavior (e.g. repeated audio, bot-like calls, spam protection)? Do you have any built-in mechanisms or recommendations to prevent the agent from continuing such calls? Can Retell help us mitigate or limit this kind of abuse on the agent level? We won’t be able to share call IDs, as recordings are removed from the Retell panel after 10 seconds and stored only in our CallPage dashboard. Thanks in advance for your help. Recording: https://drive.google.com/file/d/1ZATEB_YsmvyGqg7nQ6AkzD58dcbbWp0w/view?usp=sharing

Hey 👋 @Kate Samsonova thanks for the info — here’s a quick summary of how you can protect your AI voice agent from spam calls: Inbound Webhook filtering Enable an inbound webhook on your number Check from_number and reject suspicious calls by responding 200 without override_agent_id Detection in the agent Add logic in the prompt to detect repeated/bot-like audio and hang up quickly Abuse prevention best practices Rate limit by number or IP Use user identification if possible For public widgets, enable reCAPTCHA v3 Retell doesn’t currently auto-detect prerecorded/bot audio, so most mitigation needs to be done via webhooks, SIP provider rules, or agent logic.