ai-token-monitor vs codeburn
Menu-bar/tray app for macOS and Windows that reads Claude Code, Codex and OpenCode session logs and shows live token spend — per-model pricing with cache reads, plan-limit bars, webhook alerts. — versus — Local-first cost ledger for AI coding: reads the session files 36 tools already write and breaks every token and dollar down by task, model, project. TUI, web, desktop, menubar — no proxy, no keys.
Same job — watch AI coding spend from local session logs — at different scope: ai-token-monitor is a tray app covering Claude Code, Codex and OpenCode; CodeBurn covers 36 tools across TUI, web, desktop and menubar, adding budget guards and waste analysis.
| ai-token-monitor | codeburn | |
|---|---|---|
| Stars | 291 | 9.0k |
| Forks | 47 | 708 |
| Language | TypeScript | TypeScript |
| License | — | MIT |
| Last activity | 3 days ago | today |
| Topics | coding | coding, local |
| Curated connections | 6 | 2 |
ai-token-monitor — the curator's take
Zero-setup observability for the coding-agent bill: no API keys, no proxy — it prices the session logs your CLIs already write (cache reads included) and puts today's spend next to the clock, with live 5-hour-session and weekly plan bars that ping Discord/Slack/Telegram before you hit the wall. Offline by default; the leaderboard and chat are strictly opt-in. NOT enforcement or routing — it measures, it doesn't reduce; pair it with a gateway or harness caching fixes when the number scares you. macOS/Windows only, and note the repo currently ships no license file.
codeburn — the curator's take
The honest architecture for spend tracking — read the session files already on disk, no proxy in your API path, nothing leaves the machine — executed at the broadest scope in the category: 36 tools, four surfaces, one data source. The parts beyond pretty dashboards earn the install: waste detection with applyable (and undoable) fixes, budget guards, model comparison. Limits are inherited from the source: it sees what session logs record, and subscription flat fees vs metered API blur what a 'dollar' means — treat per-task cost as a good estimate, not accounting (pricing via LiteLLM's daily refresh). When NOT: single-tool Claude Code forensics — cc-lens and claude-code-karma go deeper there with replay and cache breakdowns.