Bug Description

Hi,

I have llm model run by vllm.

I'm using command:
hermes model
than
http://vllm:8000/v1
and this returns error: httpx.ReadError: [Errno 104] Connection reset by peer

when I set:
http://vllm:8000
than in vllm logs I see an error: "POST /chat/completions HTTP/1.1" 404 Not Found

curl from environment where agent is installed is working, for example:
curl http://vllm:8000/v1/models

the most annoying is lack of any information how final url to llm inference service looks like.

It's hard to debug and to see where is a problem.

And the biggest problem in this application is lack of service verification during setup.

So during hermes model
should be verification using some end point for example models
if there is communication, and fast feedback to the user if there are some problems
and url must be changed

Steps to Reproduce

run hermes chat
than type hi
there is error: httpx.ReadError: [Errno 104] Connection reset by peer

Expected Behavior

During installation there was sequence:

⚠ An inference provider is required for Hermes to work.

Select your inference provider:
↑/↓ Navigate Enter Select Esc Skip Ctrl+C Exit

◆ Custom OpenAI-Compatible Endpoint
Works with any API that follows OpenAI's chat completions spec
API base URL (e.g., https://api.example.com/v1): http://vllm:8000/v1
API key:
Model name (e.g., gpt-4, claude-3-opus) [anthropic/claude-opus-4.6]: llm
✓ Custom endpoint configured

so because example is: https://api.example.com/v1
than I use url with /v1 suffix

Actual Behavior

there is lack any information in vllm logs about access to the /v1/chat/completions service

Affected Component

Setup / Installation

Messaging Platform (if gateway-related)

N/A (CLI only)

Operating System

ubuntu:26.04

Python Version

bash: python: command not found

Hermes Version

0.2.0

Relevant Logs / Traceback

Root Cause Analysis (optional)

No response

Proposed Fix (optional)

No response

Are you willing to submit a PR for this?

• I'd like to fix this myself and submit a PR