agamemnon macos menu bar
token spend monitor

agamemnon

what your coding clis actually cost, priced per model, measured against the limits they really report. on disk only. no network. no telemetry.

ten clis, one helmet

Agamemnon reads local wire logs, SQLite caches, and JSONL transcripts from every coding CLI it can find. Cost is priced per model with the real cache multipliers, so a cache-heavy day reads true instead of triple. Windows are measured against the limits the CLIs actually report, never guessed, and labelled measured, estimated, or user-set. A suggestions tab flags cache thrash, wasteful model mixes, and limit pacing, all computed locally. Everything stays in ~/Library/Application Support/Agamemnon/.

claude, 3 profiles

claude code

full cache split, plus real limit hits and reset times

moonshot

kimi code

wire logs, per-turn usage records

openai

codex cli

rollout transcripts, per-turn delta

google

gemini cli

per-message token counts

sqlite

opencode

usage nested in the message store

sqlite

crush

session-level totals

github

copilot cli

output tokens only, and the card says so

activity only

cursor cli

no local token counts, so requests and lines instead of a fabricated zero

git clone https://github.com/anishfyi/agamemnon.git
cd agamemnon
./build.sh
open Agamemnon.app