Solved No transcriptions - Version 20.0 Update 10 (Build 1621 Alpha - AI 1.6.57)

[Alphabetic]

Hello!

I've had no transcriptions since 11:30 today.

Using OpenAI.

I have checked in OpenAI and we have not hit limits.

I've tried recreating API key and restarting the 3CX AI service.

Any help appreciated?

This is all i have in 3CXAI.log - verbose is logging level.

2026-09-10 16:38:47.337| Info|{} 3CX AI started v.1.6.57
2026-09-10 16:38:47.338| Info|{} Configuration loaded successfully, logging to /var/lib/3cxpbx/Data/Logs/3CXAI.log
2026-09-10 16:38:47.928| Info|{} prewarm openai properties: total_s=0.570183
2026-09-10 16:38:47.928|Debug|{} About to initialize pbx
2026-09-10 16:38:47.930| Info|{} LiveConfig: initial configuration loaded

 

[AbdullaR_3CX]

Hello @Alphabetic
If the PBX is in verbose logging mode, the transcription logs will be stored in Logs/3cxSystemService.log

 

[KyriacosS_3CX]

Hi,

Do check the logs, but also advise, this started before or after you updated to alpha2?

 

[Alphabetic]

Hi,

Do check the logs, but also advise, this started before or after you updated to alpha2?

Hi Kryiacos,

it started after Alpha 2 but this was installed about 08:30 and I had transcriptions after this.

 

[Alphabetic]

Hi all.

I managed to figure this out with the help of OpenAI (Though probably should have seen this sooner).

We ran out of API Credits at the Organisation level - I would have known this if I had access to the email the account was set up under (or checked billing but shh its a good troubleshooting exercise)

I worked this out when I tried to upload a wav to openai in Python (help of AI cos im not that clever) and it returned:



Since put additional email addresses in place for notifications

Thanks for your help everyone @IliasL_3CX

 

[KyriacosS_3CX]

Happy we could help !

 

[Alphabetic]

If anyone is interested..

Grab one of the latest wav files (copy path of one)

find /var/lib/3cxpbx/Instance1/Data/Recordings -type f -name "*.wav" -printf '%T@ %p\n' 2>/dev/null | sort -nr | head -5

Input your APIKey

read -s -p "OpenAI API Key: " OPENAI_API_KEY
echo
export OPENAI_API_KEY

Test the upload - be sure to change file path name of wav

python3 - <<'PY'
import http.client
import ssl
import uuid
import os

wav = r"/var/lib/3cxpbx/Instance1/Data/Recordings/110/<YOUR_FILE_NAME.wav"

api_key = os.environ["OPENAI_API_KEY"]

boundary = "----3CXTest" + uuid.uuid4().hex

with open(wav, "rb") as f:
    audio = f.read()

body = (
    f"--{boundary}\r\n"
    'Content-Disposition: form-data; name="model"\r\n\r\n'
    "whisper-1\r\n"
    f"--{boundary}\r\n"
    'Content-Disposition: form-data; name="file"; filename="test.wav"\r\n'
    "Content-Type: audio/wav\r\n\r\n"
).encode()

body += audio
body += f"\r\n--{boundary}--\r\n".encode()

print("WAV size:", os.path.getsize(wav), "bytes")
print("Request size:", len(body), "bytes")

ctx = ssl.create_default_context()

conn = http.client.HTTPSConnection(
    "api.openai.com",
    443,
    context=ctx,
    timeout=120
)

conn.request(
    "POST",
    "/v1/audio/transcriptions",
    body=body,
    headers={
        "Authorization": f"Bearer {api_key}",
        "Content-Type": f"multipart/form-data; boundary={boundary}",
        "Content-Length": str(len(body))
    }
)

response = conn.getresponse()

print("HTTP status:", response.status, response.reason)
print("Response:")
print(response.read().decode("utf-8", errors="replace"))

conn.close()
PY