1. Sign up free
Create an account at tools.ranomics.com. Get $5 in your wallet immediately, plus unlimited zero cost smoke runs to verify any tool works on your inputs.
Top up a USD wallet, run jobs, see the cost. New accounts start with a $5 wallet balance on signup. No subscriptions, no seats, no tiers, no minimum monthly spend.
The model is the one you already use on Modal, OpenAI, and Anthropic. You hold a USD balance, jobs debit it at the actual cost of the compute they used, and you top up when it runs low. The whole pricing surface is one number.
Create an account at tools.ranomics.com. Get $5 in your wallet immediately, plus unlimited zero cost smoke runs to verify any tool works on your inputs.
Minimum $20. Suggested amounts at checkout are $20, $50, $200, $500, and $2,500. No fixed packages. Tax is added at checkout where applicable.
Every tool form shows the estimated cost and your balance after the job. Submit, run, see the actual debit on completion. If a job ends up cheaper than estimated, the surplus returns to your wallet automatically.
Save a card to auto reload when balance drops below a threshold. Default off. Limited to one auto reload per 24 hours plus a monthly cap you set, so a runaway loop cannot drain your card.
Costs scale with the size of the run. A 5,000 design BindCraft campaign on a real target lands around $440. A single ProteinMPNN inference is closer to a nickel.
| Tool | Typical compute time | Cost per run |
|---|---|---|
| Epitope Scout | 30 seconds | under $0.10 |
| ProteinMPNN | 1 minute on L4 | around $0.05 |
| AlphaFold 2 (single fold) | 5 minutes on A100 | around $0.50 |
| ColabFold (MSA plus fold) | 8 minutes on A100 | around $0.80 |
| RFdiffusion pilot | 20 minutes on A100 | around $1.60 |
| BindCraft pilot | 60 minutes on A100 | around $4.40 |
| RFAntibody pilot | 60 minutes on A100 | around $4.40 |
| BoltzGen pilot | 30 minutes on A100 | around $3.20 |
Every job displays an estimated cost before you submit. You confirm or cancel. The estimate updates live as you change parameters.
About 40 ProteinMPNN runs. Good for first exploration after the signup balance.
One small project. Tens of folds, a single round of MPNN sequence design on a 100 backbone batch.
Multi target exploration. Run several design loops across two or three targets.
One full pilot campaign. Thousands of RFdiffusion backbones, sequence design, and structure validation on a single binder target.
Production scale. At this level a Binder Pilot service engagement is often more efficient. Talk to us.
The point of metered billing is transparency. The point of these guardrails is making sure transparency does not turn into a runaway charge while you are asleep.
Every tool has a maximum cost ceiling per job, scaled to the parameters you pick. A BindCraft run is capped at $500 even at full campaign scale. Above $1,000 estimated cost on a single job, we route you to a Binder Pilot conversation.
Your account has a daily spend cap, default $200, that you can raise or remove. Jobs that would push you over the cap are blocked until tomorrow.
Auto reload, when enabled, is limited to one charge every 24 hours plus a user set monthly cap (default $1,000). A compromised account or runaway loop cannot drain your card.
Long running jobs report progress every 15 minutes. If the actual cost runs more than 50 percent past the estimate, you get a warning email. At 100 percent over the hold, the job is killed to protect you from a bad input or an infinite loop.
It depends on the cause. Infrastructure failures on our side (Modal pod crash, tool image bug, a no progress timeout, or our preflight letting bad input through) refund the full hold. Jobs that completed normally but produced zero passing designs, or jobs you cancelled mid run, are billed for the GPU consumed. Ambiguous cases default to refund. Per tool hard caps limit how much any single run can cost.
Top ups are final. Unused balance does not expire. Start with a small top up to evaluate the platform if you are unsure.
At sustained spend above $1,000 a month, a fixed scope Binder Pilot engagement is usually more cost effective and removes the operational overhead. You get a defined deliverable, wet lab handoff if you want it, and a single point of contact for the campaign.
You top up a USD wallet in any amount from $20. Each job debits the actual cost of the compute it consumes. New accounts start with $5 in their wallet. There is no subscription, no seat license, and no fixed tier.
Yes. Auto reload is opt in and off by default. You set a balance threshold and a reload amount. To protect against runaway billing, auto reload is limited to one charge every 24 hours plus a monthly cap you control.
It depends on the cause. If the failure is on our side, like a Modal pod crash, a bug in our tool image, our preflight letting bad input through, or a no progress timeout, the full hold is refunded. If the job completed normally but produced zero passing designs, or you cancelled it mid run, you pay for the GPU consumed. Ambiguous cases default to refund. Per tool hard caps also limit how much any single run can cost in the worst case.
No. Top ups are not refundable. Unused balance remains in your wallet and does not expire. Start with a small top up if you want to evaluate the platform.
When a single job would exceed $1,000 of compute, or your monthly spend exceeds $1,000 sustained. At that scale a fixed scope Pilot engagement with wet lab handoff is usually more cost effective and removes the operational overhead.
Stripe payment processing fees are absorbed by Ranomics. Tax is added at checkout where applicable. Your wallet is credited with the exact amount you topped up.