Start free, pay as you scale
No per-seat fees, so your team can join from day one. You only pay for the compute you run, billed by the second.
Plans
Developer price: Free
For developers and small teams getting started with Polars Nanook on their own data.
- 2,000 vCPU-hours/month included, capped so you're never billed by surprise
- Optional pay-as-you-go at $0.05 / vCPU-hour
- A workspace with up to 3 concurrent clusters
- Unlimited users, plus a service account for automation
- 14 days of query history, up to 2,000 queries
- Community support on the public issue tracker
Team price: $200
/ monthFor teams running Polars Nanook in production, with unlimited clusters and support.
- 5,000 vCPU-hours/month included, worth $250 of compute
- $0.05 / vCPU-hour after that
- Unlimited users and service accounts
- 3 workspaces and unlimited concurrent clusters
- 90 days of query history (fair use)
- Support desk
- Volume discounts available
Enterprise price: Custom
For organizations that need air-gapped deployment, custom pricing or dedicated support.
- Everything in Team
- Custom compute pricing
- Unlimited workspaces and query history
- Air-gapped deployment, entirely in your infrastructure
- Dedicated Slack or Teams channel with a support SLA
- Roadmap influence
Running on a single node?
Polars is open source
Free forever, with no usage limits. Install it in seconds and run it on your own machine. When your data outgrows one node, Nanook runs the same code unchanged at scale.
Compare plans
| Feature | Developer | Team | Enterprise |
|---|---|---|---|
| Pricing & compute | |||
| Plan price | Free | $200 / month | Custom |
| Included compute | 2,000 vCPU-hours / month | 5,000 vCPU-hours / month | Custom |
| Additional compute | Optional, $0.05 / vCPU-hour | $0.05 / vCPU-hour | Custom |
| Volume discounts | Not included | Included | Custom pricing |
| Included in every plan | |||
| Polars distributed engine, no limit on cores or nodes | Included | Included | Included |
| Unified API (same code everywhere) | Included | Included | Included |
| Query profiling and built-in MCP server | Included | Included | Included |
| Deploy on AWS, Kubernetes, Ray or bare metal | Included | Included | Included |
| Organization | |||
| Workspaces | 1 | 3 | Unlimited |
| Users | Unlimited | Unlimited | Unlimited |
| Service accounts | 1 | Unlimited | Unlimited |
| Concurrent clusters | 3 | Unlimited | Unlimited |
| Query history | 14 days, up to 2,000 queries | 90 days, unlimited queries (fair use) | Unlimited |
| Support & governance | |||
| Support channel | Public issue tracker | Support desk | Dedicated Slack or Teams channel |
| Support SLA | Not included | Not included | Included |
| Air-gapped deployment | Not included | Not included | Included |
| Roadmap influence | Not included | Not included | Included |
| Start for free | Get started | Talk to us | |
Billing that follows your compute
Per-second billing
You pay only while a cluster is on, billed down to the second.
Use your committed spend
Pay through AWS Marketplace and Polars Nanook counts toward your existing AWS commitment.
Runs in your infrastructure
Compute runs next to your data, in your own infrastructure, so you're not paying to move it around.
Frequently asked questions
One virtual CPU core running for one hour. Beyond your plan's included hours it's $0.05 per vCPU-hour, metered by the second, so you never pay for more than you use.
Developer is free for trying out Polars Nanook or running small workloads, with 2,000 vCPU-hours included every month. Usage beyond the included hours is possible once a payment method is connected.
By credit card, or through AWS Marketplace to keep Nanook on your AWS bill and put it toward any committed AWS spend.
As long as a cluster is on, you pay for it, even between queries. Keep a long-running cluster or start one per job and stop it when it's done, whatever fits your workload.
The vCPU-hour price is the same wherever Nanook runs, and the same on Developer and Team. Team customers can get a volume discount, and Enterprise has custom pricing.
Team gives most teams everything they need, with 3 workspaces, unlimited clusters, 90 days of query history and a support desk. Move to Enterprise for air-gapped deployment, custom compute pricing, a dedicated Slack or Teams channel and a support SLA.