ComfyUI serves image generation to Open WebUI on the Tart VM
(ai.quineglobal.com). Models unload when idle on both sides so they
share the 64 GB.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
At 32k, opencode auto-compacted the history four times in four minutes
of reading files; after each summary the model lost the task, and after
the last it started writing tool calls as JSON text. 128k fits in ~31 GB.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Ollama LaunchAgent, derived models (mostly to set num_ctx, which
Ollama's OpenAI-compatible API won't take per request), and opencode's
config, now defaulting to qwen3-coder:30b: qwen3:32b tends to end its
turn after thinking without making the tool call it planned.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
