The Retell Community is where innovators, call center leaders, and AI enthusiasts come together to share ideas, learn best practices, and explore the future of voice automation. Join to connect with peers, get early access to product updates, and collaborate on building human-like AI voice experiences that transform customer service.

Retell AI

There was an error in mid-stream and agent hung. Is there a fallback mechanism to make this more reliable?
Call_id: `call_4ba9d69971de4f187e27ea8c26a`
https://cdn.discordapp.com/attachments/1479052688724328468/1479052689395421285/image.png?ex=69aaa2a8&is=69a95128&hm=97688cb0eadaaec4e747d4d484b7ae327bbba02732d0b371a5ce4dd89b27a1ff&

The 500 INTERNAL error originated from the upstream LLM provider

But it's eventually a part of Retell, ain't it?

I'm wondering if there is an implementation on Retell side where they invoke a fallback model (which user selects via UI) and switch to that midway in such cases

There is no user- configurable fallback LLM model that switches to mid-strean when primary model fails

LLM fallback and retries is automatically built in

So if a provider throws error (like in above case), the agent will hang?

No, the agent shouldn't just hang, in your logs, you can see the LLM error occurred twice (~17 seconds apart), which suggests Retell was retrying. If retries ultimately fail, the call would end with an error disconnection reason.