How to Make Calls with the Docker Container?

[i-am-flashh]

Hi there,

First off, thank you for your wonderful work on this project!

I'm new to this and have a basic question. I've successfully installed Docker and started the container. However, I'm not sure how to proceed with making a call. Do I need to install OpenSIPS separately and run it in parallel, or does this package itself handle calls?

Any guidance would be greatly appreciated.

Thanks!

[hasanr68]

Hi.
No expert, and struggling with similar issues. I think you need to have an Opensips run in parallel - hence the datagram connection.

[i-am-flashh]

i am running opensips docker on https://github.com/OpenSIPS/docker-opensips-cp-all-tools/

is it possible to merge these two?

[hasanr68]

Hi - that is the control panel for Opensips (not really Opensips). I am sure someone will chime in.

[Vince-0]

This project runs it's own opensips service and the AI connector service. You can see this by docker ps (in a separate session) after docker compose up (leave this session open). You will see some service logs and interactions after that. You should be able to call any extension to your server IP on port 5060 to start a call. Be aware that anyone on the internet can hit your SIP port and initiate a call so it is best to firewall it.

[i-am-flashh]

Thanks @Vince-0 for the response.

when i try to call to openai or openai@127.0.0.1 from zoiper, its been on calling state. when i end the call i get the last two lines.

ai-voice-connector-opensips  | Jan  5 07:30:56 [23] NOTICE:Started new call for B2B.230.269.1736062256.144037232
ai-voice-connector-opensips  | Jan  5 07:31:54 [23] ERROR:b2b_entities:_b2b_send_reply: Mismatch between the method in tm[8] and the method to send reply to[1]
ai-voice-connector-opensips  | Jan  5 07:31:54 [23] ERROR:b2b_entities:b2b_prescript_f: Failed to send 200 OK reply   

.env

DEEPGRAM_API_KEY=
OPENAI_API_KEY=
MI_IP=127.0.0.1
MI_PORT=8080

config.ini

ip = 127.0.0.1
port = 8080

[engine]
event_ip = 127.0.0.1
event_port = 50060

[openai]
disabled = false
key = sk-proj-xxx
model = gpt-4o-realtime-preview-2024-10-01
voice = alloy
instructions = You are a helpful assistant.
welcome_message = Hello, how can I help you?
temperature = 0.8
max_tokens = inf
turn_detection_type = server_vad
turn_detection_silence_ms = 200
turn_detection_threshold = 0.5
turn_detection_prefix_ms = 300
match = openai```

[Vince-0]

I'm not sure what's causing that. It could be a network issue with using a private loopback address. Mine is working from a public IP virtual machine.

[shazamcoder]

@Vince-0 , I too facing th following issue ! Tried with an AWS ec2 machine and my mac and windows.

All Ports were open on both machine! Getting the same error as well !

[dariusstefan]

Hi, everyone!
If you are trying to test the project with the proposed docker compose file, please make sure that you configure environment variables correctly via a well-formed config.ini file, placed in src folder, or via the .env file from docker folder.

[phamhieptel4vn]

Hi @Vince-0 ,
Can you tell me which extension to use login on Zoiper?

[Vince-0]

Hi @Vince-0 , Can you tell me which extension to use login on Zoiper?

Any extension I think, there is no authentication.

[phamhieptel4vn]

Hi @Vince-0 , Can you tell me which extension to use login on Zoiper?

Any extension I think, there is no authentication.

Thanks, i call to openai and received 1011 Error from Deepgram. I don't know if I have something wrong configured.
Log on container engine:

2025-03-14 10:28:26,238 - tid: 140601609848640 - INFO - Entering Call::get_body()
2025-03-14 10:28:26,238 - tid: 140601609848640 - INFO - Returning SDP Body:
v=0
o=FreeSWITCH 1741926586 1741926587 IN IP4 10.148.0.2
s=FreeSWITCH
c=IN IP4 10.148.0.2
t=0 0
m=audio 59909 RTP/AVP 8
a=rtpmap:8 PCMA/8000
m=video 26146 RTP/AVP 102
a=inactive
a=rtpmap:102 H264/90000
a=rtcp-fb:102 ccm fir
a=rtcp-fb:102 ccm tmmbr
a=rtcp-fb:102 nack
a=rtcp-fb:102 nack pli

2025-03-14 10:28:26,240 - tid: 140601609848640 - INFO - Starting Deepgram transcription session
2025-03-14 10:28:26,926 - tid: 140601609848640 - INFO - Deepgram STT session started successfully
ConnectionClosed in AbstractAsyncWebSocketClient._listening with code 1011: Deepgram did not receive audio data or a text message within the timeout window. See https://dpgr.am/net0001
2025-03-14 10:28:38,926 - tid: 140601609848640 - ERROR - ConnectionClosed in AbstractAsyncWebSocketClient._listening with code 1011: Deepgram did not receive audio data or a text message within the timeout window. See https://dpgr.am/net0001
send() failed - ConnectionClosed: received 1011 (internal error) Deepgram did not receive audio data or a text message within the timeout window. See https://dpgr.am/net0001; then sent 1011 (internal error) Deepgram did not receive audio data or a text message within the timeout window. See https://dpgr.am/net0001
2025-03-14 10:28:38,927 - tid: 140601609848640 - ERROR - send() failed - ConnectionClosed: received 1011 (internal error) Deepgram did not receive audio data or a text message within the timeout window. See https://dpgr.am/net0001; then sent 1011 (internal error) Deepgram did not receive audio data or a text message within the timeout window. See https://dpgr.am/net0001

[hasipfaruk]

Is there a frequency mismatch problem ?

ai-voice-connector-engine | 2025-05-29 12:34:00,511 - tid: 140392712222528 - INFO - Received message: {'type': 'SettingsApplied'}
ai-voice-connector-engine | 2025-05-29 12:34:00,595 - tid: 140392712222528 - INFO - Received message: {'type': 'ConversationText', 'role': 'assistant', 'content': 'I am calling from Health Enrollment Center under the Affordable Care Act, designed to pre-qualify individuals for government-subsidized healthcare plans with a zero-dollar premium'}
ai-voice-connector-engine | 2025-05-29 12:34:00,656 - tid: 140392712222528 - INFO - Received message: {'type': 'AgentAudioDone'}
ai-voice-connector-engine | 2025-05-29 12:34:12,513 - tid: 140392712222528 - INFO - Received message: {'type': 'Error', 'description': "We waited too long for a websocket message. Please ensure that you're sending binary messages containing user speech.", 'code': 'CLIENT_MESSAGE_TIMEOUT'}
ai-voice-connector-engine | 2025-05-29 12:34:19,195 - tid: 140392712222528 - INFO - Call B2B.484.380.1748522039.1415385181 closing
ai-voice-connector-opensips | May 29 12:34:23 [20] NOTICE:Started new call for B2B.144.307.1748522063.347989592

[hasipfaruk]

Don't forget to make the RTP with PUBLIC IP, or you won't capture any audio