Free
For trying NautAI on a real dataset, and for open research.
for up to 2 people
Start free- 3 private datasets, unlimited public
- 10 GB storage
- 5,000 synthetic rows a month
- 100 eval runs a month
- 30 days of version history
Every plan includes versioning, cleaning and evals. Paid plans add room for more data, more people and the production loop.
Annual billing saves about 20%.
For trying NautAI on a real dataset, and for open research.
for up to 2 people
Start freeFor small teams shipping their first models to production.
per seat a month, up to 10 seats, billed yearly
Start a Pro trialRecommended
For teams running the full loop, with production feedback.
per seat a month, billed yearly
Start a Team trialFor regulated data, self-hosting and large annotation programs.
annual contract
Contact sales| Feature | Free | Pro | Team | Enterprise |
|---|---|---|---|---|
| Datasets and storage | ||||
| Private datasets | 3 | Unlimited | Unlimited | Unlimited |
| Public datasets | Unlimited | Unlimited | Unlimited | Unlimited |
| Storage included | 10 GB | 200 GB | 2 TB pooled | Custom |
| Version history | 30 days | Full | Full | Custom retention |
| Workflow | ||||
| Branches and merge requests | – | |||
| Labeling queues | 1 queue | |||
| Synthetic rows per month | 5,000 | 100,000 | 1,000,000 | Custom |
| Eval runs per month | 100 | 2,000 | Unlimited | Unlimited |
| Production feedback ingestion | – | – | ||
| Managed annotation workforce | – | – | – | |
| Security and support | ||||
| Roles and permissions | – | Basic | ||
| SAML SSO and audit logs | – | – | ||
| SCIM, data residency, self-hosting | – | – | – | |
| Support | Community | Priority | Dedicated engineer, 99.9% SLA | |
Anyone who can sign in to your workspace: engineers, reviewers and annotators. Read-only viewers on Team and Enterprise are free.
Nothing stops working. Extra storage is $0.03 per GB a month, extra synthetic data is $0.80 per 1,000 rows, and extra eval runs on Pro are $0.01 each. You can set a monthly spend cap in billing settings.
Eval runs use your own model endpoints and API keys, so inference is billed by your provider. An eval run is one suite run against one dataset version.
You do. We never train our own models on your data, and you can export any dataset version, with full history, at any time.
Yes. Public datasets are free on every plan, and early-stage startups and academic labs get 50% off Pro or Team for the first year. Contact us with a few details.
Upgrades apply immediately and are prorated. Downgrades take effect at the end of your billing period; data above the new plan's limits stays read-only until you remove it.