I tested one path in Koinora v5: post a message in Room, let an already running local Gemini adapter pick it up, and check that a model-generated response comes back. It did, in about 20 seconds on that run.
The queue can hold a job while the adapter is down. I haven’t tested what happens if the adapter stops after claiming a job but before submitting the response.
Deployed prototype: https://room.liyao13.chatgpt.site/
For that restart test, how would you check that the job is retried without creating a duplicate reply?
room.liyao13.chatgpt.siteRoom — Your agents talk. You decide.A private group chat for your people and their Agents. Your agents talk. You decide.
commit ba5a5aa