Server Error (500) when trying to add a new model in self-hosted mode

[raphaelventura]

Server

• Cloud (https://app.khoj.dev)
• Self-Hosted Docker
• Self-Hosted Python package
• Self-Hosted source code

Clients

• Web browser
• Desktop/mobile app
• Obsidian
• Emacs
• WhatsApp

OS

• Windows
• macOS
• Linux
• Android
• iOS

Khoj version

1.42.10

Describe the bug

I have added llama.cpp and Ollama as AI models APIs.

Whenever I try to add a new model that is of either type, I'm getting a HTTP 500 error.

Current Behavior

The associated python trace is attached.

khoj.log

Expected Behavior

Being able to add a model through the web interface without getting this error

Reproduction Steps

• install khoj through pip as stated in the docs. I'm using uv and chose the vulkan GPU interface: CMAKE_ARGS="-DGGML_VULKAN=on" FORCE_CMAKE=1 uv pip install 'khoj[local]'
• launch khoj with USE_EMBEDDED_DB="true" khoj --anonymous-mode
• try to add a model through the web interface

Possible Workaround

No response

Additional Information

No response

Link to Discord or Github discussion

No response

[genevera]

I'm having this same issue. It's very frustrating.

[yang1002378395-cmyk]

This should be fixed by PR #1277 and #1280 which handle the case where ChatModel.friendly_name is None.

The error occurred because the __str__ method returned None instead of a string when friendly_name was null. PR #1280 provides a more robust fix with proper type handling.

Please test with the latest changes from the PR.

[raphaelventura]

Thanks, is there a way I can still install khoj using pip? From what I see in the PR's gh actions, the package is not yet published to pypi.