vibetrack
v0.1.1
Published
Track all your AI usage and spend — coding agents, LLM APIs, local models, and generative media (image/video/music/audio/3D) — in one CLI. ccusage + viberank, but universal.
Maintainers
Readme
VibeTRACKER
Track all your AI usage and spend — coding agents, LLM APIs, and generative media
(image / video / music / audio / 3D) — in one CLI. Like ccusage + viberank, but universal.
Quick start
npx vibetrack init # setup wizard (auto-detects local tools, prompts only for keys you use)
npx vibetrack sync --demo # try it: your Claude Code logs + sample data
npx vibetrack stats --domain ai # deep breakdown by provider/category/model/dayPublished to npm as vibetrack; installs the vibetracker command.
Bundled to a single file — needs only Node ≥ 18, no TypeScript, no dependencies.
Or install it globally: npm i -g vibetrack.
Track real providers
export VT_REPLICATE_TOKEN=r8_...
vibetracker connect replicate # or: vibetracker connect replicate --set token=r8_...
vibetracker sync
vibetracker total --by categoryCredentials are read from VT_* env vars or stored in ~/.vibetracker/config.json
(chmod 600). Nothing leaves your machine unless you run vibetracker upload.
Supported (12)
Higgsfield · Claude Code · Replicate · fal.ai · Runway · OpenAI · Anthropic · ElevenLabs · Luma · Kling · Suno · Udio
Trust & privacy
Every provider response and every upload passes an untrusted-data ingest boundary (ANSI/
control-byte strip, numeric bounds, prototype-pollution rejection, whitelist rebuild)
before it is stored — see docs/compliance/TRUST_MODEL.md. Uploads are self_reported by
construction; only the (upcoming) backend can mark data verified.
Commands
providers · connect <provider> [--set field=value] · sync [--demo] ·
total [--by provider|category|model|day] · upload [--handle <name>] [--url <endpoint>]
MIT
