Custom model doesn't work

[snatvb]

Hello, tried to run this:

ralphy --opencode --model opencode/glm-4.7-free "review codebase"

and it ends in 0s and says "done". With PRD it just mark everything as Done and that's all. Local models from ollama or from cloud providers doesn't matter. No errors or warnings, just marking "Done".

[Akshmit]

create opencode.json
{ "$schema": "https://opencode.ai/config.json", "permission": "allow", "model": "zai-coding-plan/glm-4.7"}

paste this inside it

then run
ralphy --opencode

[ethanwheatthin]

The above steps did not work for me. I am trying to run ralphy with opencode and using the glm-4.7-flash model through ollama

[snatvb]

create opencode.json { "$schema": "https://opencode.ai/config.json", "permission": "allow", "model": "zai-coding-plan/glm-4.7"}

paste this inside it

then run ralphy --opencode

it seems free version has been deleted, local flash ollama version doesn't work :(

[ethanwheatthin]

Would there be a way to invoke the ralphy command with claude code and ollama? https://docs.ollama.com/integrations/claude-code

[snatvb]

Would there be a way to invoke the ralphy command with claude code and ollama? https://docs.ollama.com/integrations/claude-code

I'd say opencode easier to use it

---

it seems that I had smt wrong, glm isn't free anymore, opencode/big-pickle works fine

I'd suggest to fix auto mark complete if model is wrong :)