Simple, usage-based pricing.
Start free. Paid plans include a monthly credit balance, and every metered workload draws from the same balance at $0.01 per credit.
Free
forever
Build your first forecast. No credit card.
- 500 credits / mo ($5 of usage)
- 1 workspace
- 1 seat
In every workspace
- 1 goal
- 1 shared deployment
- 1 GB ingestion / mo
Starter
per month
For solo builders shipping into production.
- 5,000 credits / mo ($50 of usage)
- 2 workspaces
- 1 seat
In every workspace
- 10 goals
- 3 shared deployments
- 20 GB ingestion / mo
Fine-tuning
Team
per month
For teams of any size shipping forecasts together.
- 50,000 credits / mo ($500 of usage)
- 5 workspaces
- 5 seats included · $99/seat after
In every workspace
- 50 goals
- 10 shared + 2 dedicated deployments
- 100 GB ingestion / mo
Fine-tuning · Dedicated hosting · Marketplace selling
Enterprise
annual contract
Mission-critical, regulated, or very large workloads.
- Custom credit allowance
- 25 workspaces
In every workspace
- 200 goals
- 50 shared + 5 dedicated deployments
- 1,000 GB ingestion / mo
Fine-tuning · Dedicated hosting · Marketplace selling · SSO / SAML · On-premises
Compare plans
Quotas are per workspace — only workspaces and seats are account-level.
| Plan | Free | StarterRecommended | Team | Enterprise |
|---|---|---|---|---|
| Pricing | ||||
| Monthly fee | $0 | $149 / mo | $2,499 / mo | Contact sales |
| Included credits / month1 credit = $0.01 · covers all compute | 500 cr · $5 | 5,000 cr · $50 | 50,000 cr · $500 | Contact sales |
| Overage rateUsage beyond included credits | Hard cap | $0.01 / credit | $0.01 / credit | $0.01 / credit |
| Your accountThe only account-level caps — everything else is per workspace. | ||||
| WorkspacesIsolated projects, each with its own full quota set | 1 | 2 | 5 | 25 |
| Seats included | 1 | 1 | 5 | Contact sales |
| Extra seats | $99 / seat / mo | Contact sales | ||
| In every workspaceEach workspace receives the full allowance shown below. | ||||
| GoalsAutonomous discovery + live forecasting | 1 | 10 | 50 | 200 |
| Shared deploymentsLive serving slots on shared infrastructure | 1 | 3 | 10 | 50 |
| Dedicated deploymentsYour own always-on serving pods | 2 | 5 | ||
| PipelinesTraining workflows | 2 | 10 | 50 | 200 |
| ModelsTrained model versions kept in the workspace | 3 | 20 | 100 | 500 |
| SegmentsTraining tables built from your signals | 3 | 25 | 200 | 1,000 |
| SignalsStored time-series values | 100K | 10M | 25M | 500M |
| Signal typesDistinct signal names (schemas) | 25 | 150 | 500 | 5,000 |
| Event streamsDistinct event keys recorded in the workspace | 1 | 5 | 20 | 100 |
| Fine-tuned adaptersPredictFM adapters trained on your data | 1 | 50 | 100 | 500 |
| BYOM modelsBring-your-own-weights uploads | 10 | 25 | 100 | |
| Monthly data allowance — per workspaceIncluded each month, in each workspace. Going over just bills credits at $0.01 each. | ||||
| Bulk ingestionImports, connectors, file uploads | 1 GB | 20 GB | 100 GB | 1,000 GB |
| Streaming eventsReal-time pushes and webhooks | 10K events | 200K events | 1M events | 10M events |
| Hot storageStored signal data, sampled daily (1 GB held for 30 days = 30 GB-days) | 30 GB-days | 300 GB-days | 1,500 GB-days | 30K GB-days |
| Dedicated hostingDays of always-on dedicated serving pods | 6 pod-days | 10 pod-days | ||
| Queries scannedData read by your analytics queries | 50 GB | 1 TB | 5 TB | 50 TB |
| Connector runsScheduled or manual connector syncs | 30 runs | 300 runs | 500 runs | 5,000 runs |
| Live goal monitoringDays of goals serving live forecasts (1 goal live for 7 days = 7 goal-days) | 7 goal-days | 35 goal-days | 100 goal-days | 300 goal-days |
| Capabilities | ||||
| Fine-tune models | Yes | Yes | Yes | |
| All model sizes | Yes | Yes | Yes | |
| Dedicated hosting | Yes | Yes | ||
| Sell on the marketplace | Yes | Yes | ||
| SSO / SAML | Yes | |||
| On-premises deployment | Yes | |||
| Get started | Get started | Get started | Talk to sales | |
Pricing questions, answered.
What's a credit?
A credit is the platform's unit of compute. Every metered workload — discovery, trainings, inference, fine-tuning, dedicated hosting — is priced in credits, so cost is predictable across tasks. 1 credit = $0.01 USD.
What does "per workspace" mean?
A workspace is an isolated project, and it's the platform's quota unit. Every quota on your plan — goals, deployments, pipelines, models, segments, signals, plus the monthly data allowances — is granted to each workspace in full. On Starter, each of its 2 workspaces gets its own 10 goals, 3 deployments, and 20 GB of monthly ingestion. Only workspaces and seats are account-level.
How do plans work?
Free is hard-capped. Starter and Team carry a monthly subscription with a bundle of included credits; anything above the bundle is overage at $0.01/credit, billed at the end of the cycle. Enterprise is a custom annual contract tailored to your requirements.
How much does a goal cost?
A goal charges credits for the work it actually does: a discovery run when drivers are found, one bounded charge per model tournament, and a small per-day charge while it's serving live. Set a monthly spend cap on any goal — autopilot pauses itself and notifies you before it goes over, not after.
What happens if a workspace goes over its data allowance?
Monthly allowances for ingestion, streaming, storage, and hosting reset each billing cycle. Eligible usage above an allowance draws from your credit balance at the published rate. Hard object caps, such as deployments per workspace, return an error instead of creating an unplanned charge.
Is there a free tier?
Yes. Free includes one workspace, 500 credits to start, and pretrained models at the small size. No credit card required.
Can I bring my own cloud?
Enterprise customers can deploy Predict.ai inside their own VPC on AWS, GCP, or Azure with a managed control plane.
Where do these numbers come from?
Plan limits and rates are loaded from the same public /v1/plans/compare configuration endpoint that the platform enforces.
Put your data to work.
Connect your own data or start with a sample. Explore automatic insights, ask questions, engineer features, and build a forecast in the same workspace.