#1185·Vane

Vane unhandledRejection: TypeError: fetch failed

Author: jdmoser1Created Aug 14, 2026Updated Aug 21, 2026

Labelsbug

Describe the bug Hello. I'm spinning up Vane for the first time and I can't get vane to communicate with a model in Ollama. When I enter a query, I get a "brainstorming" spinning circle with no progress.

I was able to enter the URL for my local Ollama server and Vane is able to get a list of models and select.

I disabled KV cache quantization and it made no difference.

(Edit to add) I tested Ollama directly and had a successful result:

podman exec ollama ollama run llama3.2 "paw patrol haiku"

To Reproduce Spin up using the compose file below

Expected behavior A successful response with a web search result

Screenshots

Additional context vane-compose.yml

services:
  ollama:
    image: docker.io/ollama/ollama:latest
    container_name: ollama
    ports:
      - "11434:11434"
    env_file: 
      - ./ollama.env
    volumes:
      - ./ollama-data:/root/.ollama
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    restart: unless-stopped

  vane:
    image: docker.io/itzcrazykns1337/vane:latest
    container_name: vane
    ports:
      - "48081:3000"
    volumes:
      - ./vane-data:/home/vane/data
      - ./vane-uploads:/home/vane/uploads

ollama.env

OLLAMA_HOST=0.0.0.0:11434
#OLLAMA_FLASH_ATTENTION=1
#OLLAMA_KV_CACHE_TYPE=q8_0

Source: ItzCrazyKns/Vane

View original on GitHub View discussion on GitHub