Multi-agent setup timeout and communication issues
I have a multi agent setup. My point of contact is backed by a cloud LLM (Kimi 2.5), and is supposed to be the coordinator, with a coder/architect backed by a local LLM (qwen3-coder:30b-a3b), and a researcher (qwen2.5:14b-instruct), both on separate Macs (m4/48GB, m1/32gb).
They're perfectly acceptable in terms of speeds, but anytime my primary agent sends a session_send, the warm up time is too long and it assumes connection issues. On top of that, the other agents see it as a user prompt instead of a session_send, so they reply as if I prompted them.
I can't find a key in the config to adjust
They're perfectly acceptable in terms of speeds, but anytime my primary agent sends a session_send, the warm up time is too long and it assumes connection issues. On top of that, the other agents see it as a user prompt instead of a session_send, so they reply as if I prompted them.
I can't find a key in the config to adjust
session_send.timeoutSeconds
