Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦
C
docs: `coli serve` carries the dashboard too, and says so for headless hosts
Mario Gutierreze064837f55fd
diff --git a/README.md b/README.mdindex 87116f36..127bcaa2 100644--- a/README.md+++ b/README.md@@ -400,8 +400,8 @@ COLI_MODEL=/nvme/glm52_i4 ./coli plan # inspect the planned VRAM/RAM/disk plCOLI_MODEL=/nvme/glm52_i4 ./coli doctor # read-only readiness checkCOLI_MODEL=/nvme/glm52_i4 ./coli doctor --deep # strict tensors/shards/index/mirror preflightCOLI_MODEL=/nvme/glm52_i4 ./coli tune # measure and save this machine's fastest safe execution profile-./coli web --model /nvme/glm52_i4 # API + web dashboard on one port-./coli serve --model /nvme/glm52_i4 # OpenAI-compatible API only+./coli web --model /nvme/glm52_i4 # API + dashboard, and opens a browser+./coli serve --model /nvme/glm52_i4 # API + dashboard, no browser (headless)```On Windows the same commands work with `python coli chat --model D:\glm52_i4`.@@ -424,9 +424,9 @@ COLI_MODEL=/nvme/glm52_i4 ./coli chat # TUICOLI_MODEL=/nvme/inkling_i4 ./coli chatCOLI_MODEL=/nvme/kimi_k3 ./coli chat-./coli web --model /nvme/inkling_i4 # API + dashboard, same port+./coli web --model /nvme/inkling_i4 # API + dashboard, opens a browser./coli web --model /nvme/kimi_k3-./coli serve --model /nvme/inkling_i4 # API only+./coli serve --model /nvme/inkling_i4 # API + dashboard, no browser```For the non-GLM engines `coli chat` starts the gateway locally and attaches the