[Bug] AI Chat fails with OpenAI-compatible provider on LM Studio (400 error & empty response) #16539

Operating system

Windows

Joplin version

3.7.18

Desktop version info

Joplin 3.7.18 (prod, win32)

Device: win32, Intel(R) Core(TM) Ultra 9 285H
Client ID: f16e86dec56e4fecb8ff99b77a612246
Sync Version: 3
Profile Version: 53
Keychain Supported: Yes
Alternative instance ID: -
Sync target: Joplin Server
Editor: Markdown

Revision: ce25482

Backup: 1.5.1
Freehand Drawing: 4.3.0
Inline TODO: 2.1.1
joplin-excalidraw-v2: 2.0.0
Rich Markdown: 0.17.1
Templates: 3.0.0

Current behaviour

I am trying to use a local large language model hosted on LM Studio via the OpenAI-compatible API.

Environment Setup:

• LM Studio running on a desktop PC in the same LAN (IP: 192.168.1.16:1234).
• Model loaded: qwen/qwen3.6-35b-a3b.
• Joplin running on a laptop.

AI Configuration:

• Chat provider: OpenAI-compatible
• Base URL: http://192.168.1.16:1234/v1
• API key: lmstudio
• Model: qwen/qwen3.6-35b-a3b

Errors:

1. Test AI configuration:
   Error. The AI provider test failed: The provider returned an empty response. Check that the base URL ends with /v1 and that a model is loaded.

2. Actual AI Chat usage:
   AI provider returned 400.

Expected behaviour

The AI Chat should work seamlessly with LM Studio's local model, allowing me to balance privacy security and AI convenience.

Logs

No response