Usage metering & billing infrastructure
Meterary vs OpenMeter
Open-source metering infrastructure for usage-based billing, self-hostable or hosted.
What OpenMeter is great at
- Real-time event ingestion and metering pipelines for usage-based products, not just AI
- Feeding usage data straight into billing systems like Stripe
- Being open source, so a team can run the whole pipeline on its own infrastructure
Where Meterary differs
- Meterlark ships a free, continuously verified AI model pricing directory out of the box — OpenMeter is metering infrastructure, not a pricing reference
- Meterlark's workspace is scoped specifically to AI token spend, with per-model input/output/cached pricing baked in, not a general metering primitive you configure yourself
- Meterlark is hosted and usable in minutes with no engineering setup; OpenMeter is typically wired into a billing stack by an engineering team
Feature by feature
Where we're not certain about a OpenMeter feature, we link out to their site rather than guess.
Free, no-signup AI model pricing directory
- OpenMeter
- —
- Meterary
- Yes
Free workload cost calculators
- OpenMeter
- —
- Meterary
- Yes
Self-hostable / open source
- OpenMeter
- Yes
- Meterary
- —
General usage-based billing/metering engine
- OpenMeter
- Yes
- Meterary
- —
Budgets & alerts scoped to AI spend
- OpenMeter
- Check their site
- Meterary
- Yes
What-if model-switching cost comparison
- OpenMeter
- —
- Meterary
- Yes
Who should pick which
OpenMeterPick OpenMeter if you're building usage-based billing for your own product and need metering infrastructure in front of Stripe or an invoicing system.
MeteraryPick Meterlark if you want to know what AI models cost right now, and track what your own AI usage is costing you without standing up metering infrastructure.
Details about OpenMeter reflect our best understanding of a product we don't operate; confirm anything that matters for your decision on their own site.
See the directory yourself.
No account needed to check a price. Sign up when you want the spend workspace.