[Bug]: Wave doesn't work with GPT-OSS (Ollama)

[w4mhi]

Current Behavior

Hi,
I want to use the wave with GPT-OSS model(s).

When I try to ask anything, I got:
Failed to post message: failed to store message: model mismatch: expected gpt-5.1, got gpt-oss:20b

Is this supported or a bug? If is supported, please could you let me know what am I missing?
Thank you!

Expected Behavior

I am expecting to work with any local model I have installed in Ollama.

Steps To Reproduce

The configuration:
{
"ollama-llama": {
"display:name": "Ollama - GPT OSS",
"display:order": 1,
"display:icon": "microchip",
"display:description": "Local GPT OSS 20B model via Ollama",
"ai:apitype": "openai-responses",
"ai:model": "gpt-oss:20b",
"ai:thinkinglevel": "medium",
"ai:endpoint": "http://localhost:11434/v1/chat/completions",
"ai:apitoken": "ollama"
}
}

Wave Version

latest

Platform

macOS

OS Version/Distribution

No response

Architecture

arm64

Anything else?

No response

Questionnaire

• I'm interested in fixing this myself but don't know where to start
• I would like to fix and I have a solution
• I don't have time to fix this right now, but maybe later

[sawka]

Testing now on a local ollama install with the gpt-oss:20b" model and it is working.

So three things:

• change your ai:apitype to "openai-chat" (/v1/chat/completions is the chat interface)
• you need to start a new chat (Cmd-K). the "wave ai" cloud chats use gpt-5.1 + responses API, so you can't continue the conversation with a different apitype/model.
• make sure to add "ai:capabilities": ["tools"] to get tools support