Pricing
Pay only for what you use. Your data, and the data of everyone who uses what you build on Genie, never reaches a third party. Ever.
Free
- ✓375 credits on signup (~375K tokens)
- ✓Genie Lite persona: fast, light
- ✓Sync + async modes
- ✓Webhook callbacks
- ✓Per-key rate limits
- ✓No card required
Pay as you go
- ✓All personas: Lite, Standard, Pro, Coder, Heavy
- ✓Priority queue: skip the line at peak load
- ✓1 credit ≈ 1,000 tokens for small personas
- ✓Configurable per-key limits + CIDR
- ✓Sync + async + webhook modes
- ✓Daily / monthly org budget caps
Enterprise
- ✓Bring-your-own runners (BYOR)
- ✓Dedicated GPU pool: no shared queue
- ✓SSO / SAML, audit log, zero-log mode
- ✓Models on YOUR infra (data sovereignty)
- ✓Custom rate limits, SLAs, support
- ✓Volume pricing on credit packs
Credit packs
Each pack adds the listed credits to your org balance. Credits never expire.
How credits work
A credit ≈ 1,000 tokensfor the smaller personas, and ~200 tokens for Heavy. You're only ever billed for completed responses. Daily and monthly budget caps can be set per-org to prevent surprises, and per-key rate limits (RPM, RPD) protect you from runaway clients.
What happens to your data
One promise sits under everything Genie does: your data never reaches a third party: not yours, and not the data of anyone using an app you build on Genie. We're a US company; your prompts are processed only on infrastructure we run, never sold, and never used to train a model.
Most inference APIs can't make that promise: under load they quietly fall back to closed-weight vendors, and your data goes wherever that vendor sends it. We have no such fallback: every model is open-weight and runs on our own GPUs, so even at peak load there is nowhere else for your data to go. That's what makes the promise auditable, and what HIPAA, FedRAMP, and any serious data-processing agreement actually require.