Everything K-culture — comebacks to K-beauty, straight to your inboxGet it in your inbox

METAL MEDIA

bekkooazar/n8meter

19HTMLMIT

A self-hosted dashboard that tells you exactly how much your n8n AI workflows cost in tokens

n8meter reads token usage from n8n workflow executions and matches it against a daily-updated catalog of over 2,500 AI model prices to calculate real costs. It shows spending broken down by workflow, client, and day, with budget alerts, all running on your own machine with no data sent to third parties. It's built for people running AI automations in n8n who have no idea what their actual API bill looks like until it arrives.

What it does

  1. Pulls token counts directly from n8n execution history (or via a webhook fallback for n8n Cloud's free tier, which lacks API access) and attributes usage down to the specific node and run
  2. Prices usage against a catalog of 2,500+ models from OpenAI, Anthropic, Google, Mistral, Bedrock and more, refreshed every 24 hours, sourced from the open LiteLLM price list
  3. Groups costs by workflow, client, or project, and can fire Slack-compatible webhook alerts when spending crosses a set budget (daily, weekly, or monthly)
  4. Runs entirely locally with data stored in a SQLite file; no telemetry or call-home, and prompts/completions are never read or stored, only token counts and costs
  5. Costs shown are list-price estimates for planning, not exact invoices — caching discounts and batch pricing aren't factored in, and estimated (non-provider-reported) token counts are flagged with a ≈ symbol

Why it matters

Anyone building AI workflows in n8n with multiple models and clients currently has to guess at costs until the provider bill lands, making it hard to catch runaway spending or price out client work. This gives a self-hosted, privacy-preserving way to see cost breakdowns in near real time without sending any workflow data to an external service.

Terms in this repo

  • n8n · 오픈소스 워크플로우 자동화 도구, AI 에이전트 파이프라인 구축에도 쓰임
  • LiteLLM price catalog · 여러 AI 모델의 가격 정보를 모아둔 공개 데이터셋
  • SQLite · 별도 서버 없이 파일 하나로 동작하는 가벼운 데이터베이스
  • webhook · 특정 이벤트가 발생하면 지정된 주소로 자동 전송되는 HTTP 요청
  • token · AI 모델이 텍스트를 처리하는 최소 단위, 요금 산정 기준이 됨

Repository description (English)

token usage per workflow, priced daily, on your own machine.

Open on GitHub

Trending repos

All repos →

Latest from METAL MEDIA