Back to browse
GitHub Repository
3 starsRust

Tokemon, a terminal dashboard to track LLM token usage

by mm65·Mar 12, 2026·1 point·0 comments

AI Analysis

●●●BangerSolve My ProblemSlickCozy

Local TUI unifies AI token costs across 16 coding tools instantly.

Strengths
  • Native parsers for 16 providers cover most major AI coding tools immediately.
  • SQLite caching ensures instant repeated runs even when logs rotate frequently.
  • MCP server integration allows AI agents to query their own usage costs.
Weaknesses
  • Relies on log file structures staying stable across tool updates.
  • Cost estimates depend on LiteLLM pricing data accuracy rather than actual bills.
Target Audience

Developers using multiple AI coding assistants

Similar To

WakaTime · OpenPipe · LangFuse

Post Description

Hey HN,

Recently I found myself using a bunch of different AI coding tools at the same time (Claude Code, Cursor, and some custom API scripts). By the end of the week, I had absolutely no idea how many tokens I was burning or what my real API costs were unless I logged into three separate billing pages.

I built tokemon to solve this for myself. It finds the log and cache files generated by these tools on your machine, parses them, handles deduplication so requests aren't double-counted, and drops the totals into a terminal dashboard. I wrote it in Rust, so it's extremely fast and performant.

The feature I end up using the most is `tokemon top`. It opens a live TUI dashboard with sparklines for daily and weekly trends. I usually just leave it running in a terminal pane as a persistent monitor since it's nice to look at and keeps my API spend top of mind.

You can grab it via cargo: cargo install tokemon

Repo is here if you want to poke around or add parsers for other tools: https://github.com/mm65x/tokemon

Let me know what you think!

Similar Projects

Developer Tools●●Solid

Agent Lens – Code assistant observability in VSCode

Parses local JSONL session files from Copilot, Claude Code, and Codex and surfaces an interactive DAG, per-agent/token metrics, and session replay directly in the activity bar — immediately useful for spotting which agents and skills actually do the work. The Claude cache-token breakdown and timeline replay are clever, concrete features that show the author dug into provider internals; adoption looks small today, but the concept and implementation give real observability where previously there was opacity.

Niche GemWizardry
pjettter
203mo ago