Setup

[frigid ice]

Make sure ollama is listening on the device IP, open the ollama port in the firewall, create your API in ollama for use in OpenClaw.

What hardware you running? I tried running some models on a 9700 XT with no success, claw sends to many tokens for the small models

[verbal path]

My hardware is pretty old. Im just experimenting with an old computer to be honest

[frigid ice]

What changes did you make? I kept getting errors about context size, tried reducing it but it still wouldn’t work

[verbal path]

I haven’t messed with context size yet

[frigid ice]

Maybe I did something wrong! Will give it another go

I ended up switched to open router gpt 5 mini or nano for simple stuff as it’s really cheap

[verbal path]

Open router is pretty cheap