The reason for latency must be something else. Can you send me your configuration like llm model, temp, tts provider and stt provider and all that
w
WadeSong0121
12/16/2024, 3:37 AM
OK,here is the config. But some user cannot even connect the service
"transcriber": {
"provider": "gladia",
"model": "fast",
"language": "en"
},
"voice": {
"provider": "openai",
"voiceId": "alloy",
"fillerInjectionEnabled": false,
"speed": 0.7
},
"model": {
"provider": "anthropic",
"model": "claude-3-5-sonnet-20241022",
"messages": [
WadeSong0121
12/16/2024, 3:39 AM
@Sahil Thanks for your help. And we think it might be with the network because some of our users are in China
s
Sahil
12/16/2024, 10:21 AM
use gpt-4o, 11labs voice and deepgram as you transcriber
Hi , Sahil, it's urgent. Appriciate for your help
I'm integrating Vapi for voice AI development and need clarification on some technical aspects since our deployment is in China:
1. Network Optimization:
1.1 Does Vapi support configuring Daily.co's geo parameter during initialization?
1.2 Which parameters in DailyAdvancedConfig can be used to optimize access from China?
1.3 Is there support for custom ICE/TURN server configuration?
2. Code example:
const dailyConfig: DailyAdvancedConfig = {
// Which specific parameters should be used for China region optimization?
}
const vapi = new Vapi('token', 'https://api.vapi.ai', dailyConfig)
3. If configuration options are limited, are there any alternative technical solutions you'd recommend?
I've attempted to set proxyUrl in DailyAdvancedConfig, but I noticed that the Daily.co session URL generated by Vapi's API doesn't reflect this proxy configuration. How can we ensure the proxy settings are properly applied to the Daily.co sessions?