### Answer Based on the available sources, I don't have specific documentation on best practices for blocking or filtering unwanted inbound callers to a Retell AI agent. What the documentation does cover related to your concern: Concurrency Limits Pay-As-You-Go users are allocated a quota of 20 concurrent calls by default. [(1) ]() This inherently provides some protection against bulk calling attacks, as calls beyond your concurrency limit won't connect. You can adjust your concurrency quota based on operational needs via the Billing page. Abuse Prevention Features Retell has added Google reCAPTCHA to Public API keys, chat, and call widgets to protect against toll fraud and spam. [(2) ]() However, this applies to widgets and public API keys rather than direct inbound phone calls. CPS (Calls Per Second) Limits You can set Calls-Per-Second caps per provider to protect deliverability and avoid carrier throttling. This is more for outbound protection but indicates rate-limiting capabilities exist. For specific guidance on blocking unwanted inbound callers or implementing caller verification, I'd recommend reaching out to: - [Contact Sales]() - [Retell AI GitHub]() Would you like me to provide a general answer based on common practices for this type of scenario?

​ To learn more, see the sources I considered: [[1] Understand concurrency & limits](https://docs.retellai.com/deploy/concurrency) [[2] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer --- ### Blocking Specific Phone Numbers for Inbound Calls Yes, you can block specific phone numbers using inbound webhooks. When a number is made public, you can set up inbound webhooks to detect and block unwanted traffic based on the incoming number. [(1) ]() ### General Abuse Prevention Practices The documentation outlines several high-level rules for preventing abuse: 1. Implement rate limiting (number-based, IP-based, etc.) for your endpoints if necessary 2. Have user identification mechanisms (KYC measures) in place if necessary 3. Have a prompt in your agent that can potentially detect unrelated calls and hang up quickly 4. Only allow functionalities or regions that you need ### For Inbound Call Protection Specifically The recommended approach is to configure your inbound webhook to inspect the caller's phone number and decide whether to accept or reject the call before it connects. This allows you to maintain a blocklist or implement logic to identify suspicious calling patterns. For widget-based interactions (chat/callback) , you can enable Google reCAPTCHA v3 protection to prevent bots from abusing your endpoints. [(1) ]() [(2) ]() ---

​ To learn more, see the sources I considered: [[1] Prevent abuse](https://docs.retellai.com/reliability/prevent-abuse) [[2] Public Keys](https://docs.retellai.com/accounts/public-keys) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer Based on the available sources, I don't have specific documentation on a built-in feature for Retell AI agents to automatically detect when they're speaking with another AI voice agent. However, from the previous thread context, there are some approaches you could consider: Prompt-Based Detection You can include instructions in your agent's prompt to potentially detect unrelated or suspicious calls and hang up quickly. This could include prompting the agent to listen for patterns typical of AI-generated speech (repetitive phrasing, unusual cadence, lack of natural conversational cues) . Voicemail Detection as a Reference Retell AI does have

enable_voicemail_detection

functionality that detects whether a call enters a voicemail. [(1) ]() This shows the platform has some capability for detecting non-human audio patterns, though this is specifically for voicemail, not other AI agents. For a robust solution to identify AI-to-AI calls, you'd likely need to implement custom logic through: - Webhook-based caller analysis - Prompt engineering to challenge callers with questions that might trip up other AI systems - Behavioral pattern detection in your backend For more specific guidance on this use case, I'd recommend reaching out to: - [Contact Sales]() - [Retell AI GitHub]()

​ To learn more, see the sources I considered: [[1] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.