Show HN: FiscaLLM:open-source per-user AI cost for subscription apps

FiscaLLM, an MIT-licensed open-source tool published to npm and PyPI as the `fiscallm` package, tracks per-user AI spend for subscription apps by reading only the model name and token counts from each AI response and joining that cost against revenue from RevenueCat or Stripe webhooks on the user ID. The tool prices cost on arrival — exact `costUsd` when supplied, otherwise per-token rates for input, output, cache read and cache write or per-unit rates for images, audio and video — and flags unknown models as unpriced rather than counting them as zero. It adds per-user budgets with email, Slack or signed-webhook alerts and optional spending caps via `meter.check(userId)`, which fails open if FiscaLLM is slow or down.

See the AI cost of every single user. FiscaLLM adds up what each user costs you in AI and sets it next to what they pay you, so you can see who is profitable and who is losing you money. Open source MIT and free. Use the hosted service at fiscallm.com https://fiscallm.com free under fair use, best effort , or run your own https://github.com/magna56/fiscallm/blob/main/SELF-HOSTING.md . Live demo no sign-up https://fiscallm.com/demo · Docs https://fiscallm.com/docs · How it works https://fiscallm.com/how-it-works · What it covers https://fiscallm.com/coverage · Accuracy https://github.com/magna56/fiscallm/blob/main/ACCURACY.md · Roadmap https://github.com/magna56/fiscallm/blob/main/ROADMAP.md board https://github.com/users/magna56/projects/1 · Changelog https://github.com/magna56/fiscallm/blob/main/CHANGELOG.md · Self-hosting https://github.com/magna56/fiscallm/blob/main/SELF-HOSTING.md · Contributing https://github.com/magna56/fiscallm/blob/main/CONTRIBUTING.md Sample data from the demo, not a real customer. 1. Create an app at fiscallm.com https://fiscallm.com or your own copy and copy the ingest key. 2. Add one line after each AI call it reads only the model name and token counts, never prompts or answers : js import { createMeter } from "fiscallm"; // npm install fiscallm Python: pip install fiscallm const meter = createMeter { apiKey: process.env.FISCALLM KEY, waitUntil } ; const res = await openai.chat.completions.create { model, messages } ; meter.trackOpenAI res, { userId, feature: "chat" } ; 3. Connect revenue : paste a RevenueCat or Stripe webhook URL, or send payments yourself with the revenue API. No AI key handy? FISCALLM KEY=fsk ... node examples/send-sample-data.mjs fills a test app with sample users so you can see it working. AI apps charge a flat price but pay a variable cost. One user costs $0.20 a month in AI, another costs $25. Your AI provider's bill shows tokens, your subscription tool shows payments, and neither shows the user. js import { createMeter } from "fiscallm"; const meter = createMeter { apiKey: process.env.FISCALLM KEY, waitUntil } ; const res = await openai.chat.completions.create { model, messages } ; meter.trackOpenAI res, { userId, feature: "chat" } ; python from fiscallm import Meter meter = Meter api key=os.environ "FISCALLM KEY" res = client.chat.completions.create model=model, messages=messages meter.track openai res, user id=user id, feature="chat" - The SDK reads only the model name and token counts from the response. Never prompts or answers. - Events are batched and sent in the background. Tracking never throws and never blocks your response. - Each event has a client-generated UUID, so retries cannot double count. - Revenue comes from a RevenueCat webhook, or from your own code with trackRevenue / track revenue . - Both are joined on the user ID. RevenueCat aliases are resolved, so anonymous and logged-in users merge. - Budgets and alerts : email, Slack or signed webhooks when a user goes over budget, when a paying user loses money, or when one user burns a huge number of tokens in 5 minutes a runaway loop . - Spending caps : mark a per-user budget as Block and call meter.check userId before an AI call TypeScript, Python . It says no once the user reaches the amount, and fails open if FiscaLLM is slow or down. Cost is priced on arrival: exact costUsd if given, otherwise per-token input, output, cache read and cache write priced separately or per-unit images, audio, video from a price table. Unknown models are flagged as unpriced, never counted as zero. Net revenue is price × 1 − tax − store commission . | Path | What it is | |---|---| | apps/web | Next.js site and dashboard | | apps/api | Hono API: ingest, RevenueCat webhook, revenue API, cron jobs | | packages/shared | Drizzle schema, pricing, metrics, alerts | | packages/sdk | TypeScript SDK fiscallm on npm | | packages/sdk-python | Python SDK fiscallm on PyPI | Stack: TypeScript, Next.js, Hono, Postgres Supabase with Drizzle, hosted on Vercel. - It needs a backend between your app and the AI provider. Apps that call a provider straight from the phone are not covered yet. - Revenue comes from RevenueCat, Stripe beta, USD only or the revenue API. Direct App Store and Google Play connections are not built yet. - No AI key needed: FISCALLM KEY=fsk ... node examples/send-sample-data.mjs fills the dashboard with sample users one heavy user who loses money . - Copy-paste starting points for Next.js https://github.com/magna56/fiscallm/blob/main/examples/nextjs and FastAPI https://github.com/magna56/fiscallm/blob/main/examples/fastapi , see examples https://github.com/magna56/fiscallm/blob/main/examples . - Run your own copy: SELF-HOSTING.md https://github.com/magna56/fiscallm/blob/main/SELF-HOSTING.md . pnpm install pnpm dev:web http://localhost:3000 pnpm dev:api http://localhost:8787/api/health pnpm test Copy .env.example to apps/web/.env.local and packages/shared/.env and fill in the values. pnpm db:generate pnpm db:migrate Issues and pull requests are welcome. Read CONTRIBUTING.md https://github.com/magna56/fiscallm/blob/main/CONTRIBUTING.md first it includes a short contributor agreement . The plan, including what we only build if there is demand, is in ROADMAP.md https://github.com/magna56/fiscallm/blob/main/ROADMAP.md , with a public board https://github.com/users/magna56/projects/1 . Vote with a 👍 on roadmap issues https://github.com/magna56/fiscallm/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap or share ideas in Discussions https://github.com/magna56/fiscallm/discussions/categories/ideas . Where help is most wanted: prices for more models and providers data only, no code knowledge needed , connectors for other billing systems Adapty, Superwall , examples for AI frameworks, and a comparison of computed costs with a real provider invoice. See the good first issues https://github.com/magna56/fiscallm/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22 . Built and maintained by Mayur https://www.linkedin.com/in/mayur-agnani/ as an independent, open-source side project. Questions, bugs and ideas: open an issue https://github.com/magna56/fiscallm/issues or email theaicommit@gmail.com mailto:theaicommit@gmail.com . Thank you to everyone who helps. Contribute https://github.com/magna56/fiscallm/blob/main/CONTRIBUTING.md : a good place to start is an issue labelled good first issue . MIT. See LICENSE https://github.com/magna56/fiscallm/blob/main/LICENSE .