On Intel macbook -- browseros server not connecting #318
Description
Issue Type
Agent Issue
Operating System
macOS
Description of the bug
“Failed to fetch (127.0.0.1: 9100)” when testing any LLM provider on macOS
Steps to Reproduce
BrowserOS is up to date
BrowserOS v0.36.0_x64 / v0.36.1_x64 / v0.36.2_x64
Версия: 142.0.7552.49 (Официальная сборка) (x86_64)
Screenshots / Videos
Hi, I’m running BrowserOS on macOS (Intel) and I’m unable to use the “Test” button for any custom LLM provider.
Steps to reproduce:
Open Settings → BrowserOS AI.
Add a provider (tested with Gemini and OpenRouter).
Use the following settings:
Gemini
Base URL: https://generativelanguage.googleapis.com/v1beta
Model: gemini-2.5-flash
OpenRouter
Base URL: https://openrouter.ai/api/v1
Model: openai/gpt-4o
Enter a valid API key and click Test.
Expected behavior:
The provider test succeeds and confirms the connection.
Actual behavior:
The test fails immediately with the error:
Failed to fetch (127.0.0.1: 9100)
The same error appears for different providers, so it looks like the internal local server/proxy on 127.0.0.1:9201 is not reachable.
Additional info:
OS: macOS (Intel)
Elasticsearch is running successfully on http://localhost:9200 and responds correctly.
Restarting BrowserOS and changing providers does not fix the issue.
Screenshots attached show the provider configuration and the error message.
Could you please check why the internal test uses 127.0.0.1: 9100 and how to fix or bypass this on macOS?
BrowserOS Version
No response
Additional Context
No response