[bug] Stuck on scanning and then exit

[as8ASd3]

Command that failed

ptai start http://localhost:3000/

pentest-ai version

ptai --version ptai 0.14.0

Python version

python --version Python 3.13.13

OS

Kali GNU/Linux 2024.2

Expected behavior

`└─$ ptai start http://localhost:3000/
pentest-ai performs real network operations. Authorized targets only.
AUP: https://pentestai.xyz/aup · Terms: https://pentestai.xyz/terms

╭───── Starting Engagement ──────╮
│ pentest-ai v0.14.0 │
│ Target: http://localhost:3000/ │
│ Scope: full │
│ Intensity: normal │
╰────────────────────────────────╯
agent_mode: 274 action handlers registered
⠹ Scanning http://localhost:3000/...
`

Actual behavior

ptai start http://localhost:3000/
pentest-ai performs real network operations. Authorized targets only.
AUP: https://pentestai.xyz/aup  ·  Terms: https://pentestai.xyz/terms

╭───── Starting Engagement ──────╮
│ pentest-ai v0.14.0             │
│ Target: http://localhost:3000/ │
│ Scope: full                    │
│ Intensity: normal              │
╰────────────────────────────────╯
agent_mode: 274 action handlers registered

⠙ Scanning http://192.168.184.128:3000...

then nothing happens and then exit

Steps to reproduce

1.export OLLAMA_HOST=http://OLLAMA_HOST:11434
2.ptai start http://localhost:3000/

Logs and config

[poeylizn]

I encountered the same problem, and I also tried to find the report in the ~/.pentest-ai path, but there was nothing there. Have you found the cause and the solution?

[poeylizn]

[qusaialhaddad]

facing the same , any fix for that

[as8ASd3]

problem is still occurring after code change

[poeylizn]

i only set the OPENAI_API_KEY

[poeylizn]

i used the 0.16.1 depend on DeepSeek deployed in Cloud，but the error is

i guess that the reason is the pentest-ai try to connect my local model ？

if it is，how can i setup my model‘s BASE_URI and MODEL_NAME ?

[poeylizn]

After setting up these environment variables, the LLM call failed with a 400 status code. How can I view the logs?

[poeylizn]

i solved this problem by upgraded to v0.16.2

[0xSteph]

Glad 0.16.2 sorted it for you, and sorry about the back and forth on this thread.

One thing worth flagging: your "400 with no info" complaint pointed at a real gap I missed. The providers were dropping the upstream response body on 4xx, so even when DeepSeek (or others) returned an actionable error, the user only saw "Bad Request" with no reason. v0.16.3 fixes that across all four providers. Now you'd see the actual error message from the upstream embedded in the exception, not just the status code.

So your report ended up shaking loose another fix for the next person who runs into something like this. Thanks for sticking with the thread. Pushing 0.16.3 right now.