Step 1 · Detect your hardware
LeviathanTalon scales from a single agent on an 8GB laptop to 2,832 on a server. Click DETECT and pick the recommended profile, or override below.
Hardware-aware sizing
Click DETECT to autoconfigure based on what you have. Or pick a profile below.
Step 2 · One model on all 708 LLM slots (optional · profile may have done this)
Or pick a single provider + model and pour it into every slot. Kimi K2.6 1T via OpenRouter, Llama 3.3 70B via Ollama, anything OpenAI-compatible.
Fill all 8 slots
When the same config covers multiple slots, the swarm uses ONE provider instance backing N positions — efficient, cheap.
Step 3 · Adjustable scale
Tune concurrency. Lower LLM-per-sub for smaller machines. Crank the binary slider for execution throughput.
LLM per sub
1..100 — 7 subs × this = LLM workers
1..100 — 7 subs × this = LLM workers
100
Binary pool
0..2,124 — ZeroClaw concurrency
0..2,124 — ZeroClaw concurrency
2,124
Total concurrent capacity
2,832 agents
Step 4 · Per-slot overrides (optional)
Skip this if Step 1 or 2 already configured every slot. Otherwise mix providers freely.