[Feature]: Support Local IP Address

### Problem or Motivation

## Summary

Allow OpenMAIC to connect to a locally hosted (for example) LM Studio instance using a private/local network IP address.

---

## Environment fir private use

- **PC A**: Running OpenMAIC
- **PC B**: Running LM Studio and other local generative AI services
- Both machines are in the same local network (LAN)

---

## Current Behavior

LM Studio is running on PC B and is accessible via a local IP address, e.g.:

`http://192.168.1.25:1234/v1`

When configuring OpenMAIC on PC A to use this endpoint, the request is rejected with:

`Local/private network URLs are not allowed`

As a result, OpenMAIC cannot connect to LM Studio or other locally hosted AI services running on another machine in the same network.

---

## Expected Behavior

OpenMAIC should allow:

- Connections to private IP ranges:
- 192.168.x.x
- 10.x.x.x
- 172.16.x.x – 172.31.x.x
- Custom base URLs pointing to LAN-hosted model endpoints
- A configurable option to explicitly allow private network access

---

## Use Case

Typical local development setup:

**PC B (AI Server)**
- LM Studio
- Other local LLM backends
- Dedicated GPU

**PC A (Client / Docker Machine)**
- OpenMAIC

This setup enables:

- GPU isolation
- Better hardware utilization
- Clean separation between inference server and client
- Scalable local AI infrastructure

Currently, this architecture is blocked due to private network URL restrictions.

---

## Thanks for reading
Thank you for taking the time to review this request.
It would be greatly appreciated if this feature could be considered and implemented.

### Proposed Solution

One of the following approaches would solve the issue:

1. Add a configuration option to allow private network URLs
2. Add an environment variable such as:

`ALLOW_PRIVATE_NETWORK=true`

### Alternatives Considered

_No response_

### Area

Classroom generation

### Additional Context

_No response_

Locate the validation that emits "Local/private network URLs are not allowed" and trace how custom model base URLs are configured. Review the security implications of allowing private ranges and compare an explicit configuration option with the proposed environment variable. Done means LAN-hosted endpoints can be enabled intentionally, while the default restriction and coverage for the listed private ranges are tested.

Written by the indexing model from the issue text.