#Clawdbot gets errors trying to use LMStudio.

1 messages · Page 1 of 1 (latest)

knotty elk
#

Got you! Here's a clean bug report:

bug Bug Report: LMStudio Integration Crashes with "Cannot read properties of undefined (reading 'includes')"

Environment:

  • Clawdbot running on AWS
  • LMStudio on local machine via Tailscale
  • Model: Qwen 3 14B (qwen/qwen3-14b)
  • Channel: Discord

What's happening:
When trying to use an LMStudio model, Clawdbot successfully makes the API call (completes in ~67ms), but crashes during response processing with:
Cannot read properties of undefined (reading 'includes')

#

Config (simplified):
"models": {
"providers": {
"lmstudio": {
"baseUrl": "http://100.70.114.84:1234/v1",
"apiKey": "lm-studio",
"api": "openai-completions",
"models": [
{
"id": "qwen/qwen3-14b",
"name": "Qwen 3 14B",
"contextWindow": 32000,
"maxTokens": 8192,
"input": ["text"]
}

#

]
}
}
}

What works:

  • white_check_mark LMStudio API responds correctly to direct curl requests
  • white_check_mark Clawdbot starts and recognizes the model
  • white_check_mark Claude fallback works perfectly

Root cause (suspected):
The error happens in Discord message formatting code, likely checking model capabilities with something like model.input.includes('image') but the property is undefined. Missing optional chaining (?.) or null check.

Steps to reproduce:

  1. Configure LMStudio provider with a custom model
#
  1. Send a Discord message to trigger the model
  2. Observe crash after API call completes
#

Clawdbot gets errors trying to use LMStudio.

knotty elk
#

@mortal kelp