Put two or more machines on your network and they share one scan queue: while one machine reads your code, another can write up the report, and every model call goes to the machine best placed to answer it. A call is sent to another machine only when that machine can hold the whole prompt, so the model sees everything it was asked about. A machine working through a long call is never mistaken for one that has gone quiet, and the scan it is running stays on it. Each machine is measured by the calls it answers itself, so a fast machine never slows down because a slower one shares the load.
The fleet is also easy to see and keep running. The Scan Queue shows which machine is running each call, how long that call has been going and when that machine last answered. When you update Ghost Architect™ Local, running scans pause, wait for the new version on every machine, and carry on where they stopped, and the daily report counts them as paused, not cancelled. The result is more work from the hardware you already have, and a clear view of where it is going.
Ghost Architect™ Local is out of beta and commercially released.
If you want an evaluation trial, the download page tells you how to get the tarball, verify it, and start the evaluation trial on your own hardware.
Request Ghost Architect™ LocalEJ Wisner, founder of Ghost Architect™. Local pricing is monthly, per team size, with no per-token bill; the numbers are on the pricing page.