Continue 0.8.xModel / API

Continue: model not responding / request hangs forever

FetchError: socket hang up (local model)
Affected versions
0.8.x
Fixed in
Last verified
Continue 0.8.52

Root cause

The configured local model endpoint (Ollama/LM Studio) is not reachable at the URL/port in config.json, so requests hang until timeout.

How to reproduce

Set a local model in config.json, send a chat; the spinner never resolves.

Solutions (2)

Ordered by community votes. Always check the verified version before applying.

44
#1Applies to 0.8.x

Confirm the local server is up: `curl http://localhost:11434/api/tags` for Ollama. Match the `apiBase` in config.json exactly (host, port, trailing slash).

Last verified: Continue 0.8.52 · community

21
#2Applies to 0.8.x

Pull the model first (`ollama pull qwen2.5-coder`) — a missing model causes the request to stall rather than error clearly.

Last verified: Continue 0.8.50 · community

Related errors