ZenML pricing guide details
Pricing Overview
ZenML has three plans: a free, self-hosted Open Source plan, a hosted Starter plan at $39 a month, and a custom-priced Enterprise plan. Open Source covers ZenML and Kitaru under Apache 2.0, with pipelines, traces, evals and replay, unlimited pipeline runs and team members, and community support, all run on your own infrastructure.
Starter is one plan at one price: $39 a month for 1 AI agent and 1 ML project with up to 100 pipeline runs a month, and unlimited seats. It adds a hosted AI engineer that watches your agents and ML pipelines, plus schedules, triggers and alerts, and starts with a 14-day trial that needs no credit card. ZenML counts an AI agent as anything you want the AI engineer to watch: its code, in one repo or several, and the traces it sends.
Enterprise covers more agents, ML projects, pipeline runs and workspaces, priced on what you run, and adds on-prem or air-gapped deployment, your own LLM endpoint, SSO, SCIM, custom roles, audit logs, an SLA and a named engineer.
Plan Comparison
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Open Source | Starter | Enterprise | |
|---|---|---|---|
| Price | Free | $39/month | Custom |
| Hosting | Self-hosted | SaaS | Your cloud or on-prem, air-gapped if needed |
| Free trial | — | 14 days, no credit card | — |
| AI agents | Self-hosted | 1 | Custom |
| ML projects | Self-hosted | 1 | Custom |
| Pipeline runs a month | Unlimited | 100 | Custom |
| Team members | Unlimited | Unlimited | Unlimited |
| SSO, SCIM, custom roles, audit logs | — | — | Included |
| Support | Community | Dedicated + SLA |
The detail that matters most: Open Source already allows unlimited pipeline runs, while Starter includes 100 runs a month for one ML project. Paying for Starter buys the hosted AI engineer with schedules, triggers and alerts, not more capacity than the free framework has.
Hidden Costs and Considerations
- Compute is always separate. ZenML is a metadata layer on top of your own infrastructure: data, artifacts and compute stay in your VPC. Cloud compute, GPUs, storage and networking are billed by your cloud provider on every plan, and for a busy training workload they can cost more than the ZenML plan.
- Starter is sized by agents, projects and runs. It covers 1 AI agent, 1 ML project and up to 100 pipeline runs a month. More agents, projects or runs mean an Enterprise contract, priced on what you run.
- Model costs stay on your bill. Replays run with your own model keys, so those model costs stay on your bill.
- Governance is Enterprise-only. SSO, SCIM, custom roles, audit logs and on-prem or air-gapped deployment are Enterprise features. Separately, ZenML holds a SOC 2 Type II report and ISO 27001 certification.
Cost Estimates by Team Size
- Solo engineer or small team experimenting: $0 on Open Source, plus your own infrastructure and the time to run and upgrade the server yourself.
- Team of 5 with one AI agent and one ML project: Starter at $39/month, if 100 pipeline runs a month are enough. Seats are not charged, so the plan price is the same for 5 people or 20.
- Team of 20: the same $39/month while the team stays on one AI agent and one ML project within 100 runs a month; more agents, projects or runs mean an Enterprise contract.
- Regulated organisation: Enterprise, priced on request, for SSO, audit logs and air-gapped deployment.
How ZenML Pricing Compares
Weights & Biases lists a free tier and Pro at $60/mo, with Enterprise on request. Flyte is open source and free, and its managed offering through Union.ai has a Team plan at $950/month that includes $950 of monthly usage credit. Amazon SageMaker is pay-as-you-go by component.
Two things set ZenML apart. Its paid plan is priced by AI agents and ML projects rather than by seats, which keeps the bill flat as a team grows. And compute is never included, so the $39/month pays for the hosted AI engineer alone — a team comparing it with Union.ai's Team plan should note that the $950 there comes with usage credit.