Start Free. Scale As You Grow

A Noveum Credit is a unified unit of compute that powers all LLM activities on the platform. 1 credit = 1 premium LLM scorer check on up to 4,000 input tokens. Other activities burn credits at different rates: 2 credits per voice audio scorer check, except Speaking Over User at 2 credits per started audio minute, 100 credits per NovaSynth voice call minute, 12 credits per text conversation, and 8 credits per NovaPilot trace analyzed with a minimum of 800 per report. Scorers labeled Free cost 0 credits.

The Free plan includes 2,500 credits per month: enough for approximately 25 minutes of voice agent testing, 208 text conversation tests, ~312 NovaPilot trace analyses, or 2,500 LLM eval scorer-items. No credit card required to sign up.

Monthly plan credits reset at the start of each billing cycle and do not roll over. However, add-on credit packs that you purchase separately never expire: they accumulate in your credit wallet and are spent after your monthly plan credits are exhausted.

Credit-consuming work hard-stops when the enforced credit limit is reached: the UI shows a quota message and affected API calls return 429 QUOTA_EXCEEDED. Buy a one-time top-up or upgrade the plan before continuing.

Yes! NovaSynth voice and text testing is available on every plan including Free. It is credit-gated, not plan-gated: meaning Free users get approximately 25 minutes of voice testing or 208 text conversations from their 2,500 monthly credits. When credits run out, you can buy an add-on pack or upgrade your plan.

NovaPilot recommendation reports are available on all plans including Free. It costs 8 credits per analyzed trace with a minimum of 800 credits per report, so your monthly credit allowance determines how many traces you can analyze: approximately 312 on Free, 1,250 on Pro, and 3,125 on Growth.

The published library labels 30 scorers Free and 82 Premium. The label is explicit metadata: some deterministic or signal-based scorers are Premium. Free scorers cost 0 credits. Premium text scorers cost 1 credit per scorer and eligible dataset item for every 4,000 estimated input tokens, rounded up. Premium audio scorers cost 2 credits per check, except Speaking Over User at 2 credits per started audio minute.

Yes. You can upgrade your plan immediately from the billing settings in your dashboard: new limits take effect instantly. Downgrades take effect at the start of your next billing cycle. Annual plans can be upgraded at any time; downgrading an annual plan to a lower tier will be prorated.

Yes: paying annually saves you approximately 2 months (~17% off). For example, Pro is $690/year vs $828/year when paid monthly (save $138), Growth is $990/year vs $1,188/year (save $198), and Max is $1,990/year vs $2,388/year (save $398). The Free and Enterprise plans are not affected by billing cadence.

No. The current implementation hard-stops credit-consuming work and span ingestion at the enforced limits rather than accruing overage charges. Buy a credit top-up where available or upgrade the plan before continuing.

A trace is one recorded request-response lifecycle through your AI application: for example, a single LLM call, one RAG pipeline execution, or one agent turn. The Python SDK or a connected AI Gateway records the operations you instrument or route through it. Multi-step agent workflows can contain many spans but count as one trace.

Enterprise pricing and enforced usage limits are customized by contract. Managed cloud, BYO ClickHouse, on-premise, and Kubernetes/Helm deployment options can be discussed alongside support, service levels, and retention requirements. Contact us for the terms that apply to your deployment.