ringd

ringd is a tool designed to provide authentication for AI voice agents, focusing on verifying the identity of callers before sensitive actions are performed. It addresses the challenge of ensuring that automated voice agents only execute tasks such as booking appointments, processing payments, or making account changes after confirming the caller's identity. The tool is positioned as a solution for developers who need to integrate identity checks into their voice AI applications without adding unnecessary complexity to their codebase.

Key features of ringd include the ability to add authentication checks with minimal code changes, exemplified by a decorator-based approach that allows developers to specify verification requirements for specific actions. The platform offers fine-grained configuration for each verification flow, enabling developers to tailor identity checks to the sensitivity of the requested action. For routine requests, verification can remain minimal, while actions requiring higher confidence can trigger additional steps such as voiceprinting, SMS, or email verification.

ringd handles common issues in caller authentication, including shared numbers, spoofed caller IDs, missing numbers, duplicate records, and mismatched emails, ensuring that these potential problems are addressed before the AI agent proceeds. The tool also maintains a complete history of every verification attempt and decision, which can be exported for security review or audit purposes.

Integration with existing technology stacks is supported, with references to compatibility with platforms such as Supabase, Neon, Prisma, Attio, and Airtable.