Retell AI offers a range of customization options, including speaking rate, voice temperature, and ambient sound addition. It can handle interruptions and speech isolation, with an end-to-end latency of 800ms. The product is usage-based, priced at $0.1-0.17 per minute, and primarily targets developers through its API. However, it can also be tested without coding via their dashboard. The company has invited feedback from the community and is excited to see the applications users will create with their API.
Key takeaways:
- Retell AI is a conversational speech engine designed to help developers build natural-sounding voice AI, abstracting away the complexities of AI voice conversations.
- Retell AI follows the paradigm of having speech-to-text, LLM, and text-to-speech components, but adds additional conversation models to orchestrate the conversation while allowing maximum configurability for developers.
- Retell AI can achieve 800ms end-to-end latency, handle interruptions, speech isolation, with customization options like speaking rate, voice temperature, and adding ambient sound.
- Retell AI's main product is a developer-facing API, but it can be tried without writing code via their dashboard. The product is usage-based and costs $0.1-0.17/min.