Usage

How I work with AI.

I keep frontier models wired into my CLI, my editor, and my agents — Claude for reasoning, GLM for speed, Kimi for breadth. Skills and docs keep them coherent; routing and caching keep them cheap. Below is everything I've actually pulled through the pipe — tokens, requests, cache hits, model mix — straight from the UMANS API.

Updated moments ago

Daily tokens

Tokens consumed per day — last 7 days.

Tools I build with

Every day

Claude
Codex
Kimi
GLM
OMP
UMANS

—

Total tokens

Last 7 days

—

Requests

Across all providers

—

Cached tokens

Prompt cache reads

—

Plan

—

Model breakdown

Per-model token usage — last 7 days.

Model Requests Tokens in Tokens out Cached Total
Loading…