Both help you see what AI costs per user. The difference is what they compare it with, what they cost, and where your data lives. Checked on 2026-11-08 from Portkey's public page; we make FiscaLLM, so tell us if anything about Portkey is out of date.
| FiscaLLM | Portkey | |
|---|---|---|
| What it is | AI cost per user joined to real revenue: margin per user, plan and free trial | AI gateway with observability, guardrails, prompt management and budgets on virtual keys |
| Price | Free. Hosted with fair use (100,000 events per app per month), or self-hosted with no limit | Developer free (10k logs a month), Production $49 a month (100k logs), Enterprise custom |
| Hosting | Hosted or self-hosted | Hosted; the open-source gateway can be self-hosted; Enterprise adds VPC |
| Open source | Yes, MIT (server, dashboard and SDKs) | The gateway is MIT; the hosted platform is not |
| Revenue used for margin | Real revenue from RevenueCat (one-click connect and import), Stripe (webhook and 12-month import), CSV upload, or a revenue API | Logs cost with custom metadata such as user, but does not import subscription revenue to compute margin |
| Stripe | Per end user: payments, trials, cancellations, refunds and lost disputes from a webhook, 12 months of history, Stripe's exact fee and any currency with a read-only key | No |
Pick Portkey if you want a hosted AI gateway with guardrails and routing. FiscaLLM needs no gateway and adds revenue and margin per user.
Open the demo dashboard (sample data, no sign-up), read the docs, or start free. All comparisons: AI cost per user tools compared.
Yes. FiscaLLM is free and open source (MIT): use the hosted service at fiscallm.com or run your own copy. Portkey: Developer free (10k logs a month), Production $49 a month (100k logs), Enterprise custom.
Yes. FiscaLLM only needs one tracking call after each AI request and a revenue source, so it runs next to other tools.