Open Source
Self-hostedRun it yourself
Free
Your infrastructure · Your limits
Includes
- ZenML and Kitaru, Apache 2.0
- Driven by your own coding agent
- Pipelines, traces, evals and replay
- Community support
ZenML Pro pricing
ZenML Pro, $39 a month
ZenML Pro is our paid platform. It hosts Kitaru and ZenML for your team and adds a team of AI engineers that watch your AI agents and ML pipelines around the clock. Starter covers one AI agent, one ML project and your whole team. Need more, or on-prem? Talk to us.
An AI agent is anything you want your AI engineers to watch: its code, in one repo or several, and the traces it sends. Starter covers one AI agent and one ML project. Need more? Talk to us.
Pro Starter
SaaSFor startups shipping their first agent
$39/month
1 AI agent · 1 ML project · Unlimited seats
Everything in Open Source, plus
- AI engineers for your agents and ML pipelines
- Finds what your evals miss, on its own
- Schedules, triggers and alerts
- Fully hosted, set up in minutes
Pro Enterprise
Your cloud or on-premFor more agents, more projects or regulated data
Custom
More agents · More projects · More workspaces
Everything in Starter, plus
- Multi-workspace setup
- On-prem or air-gapped
- Your own LLM endpoint
- SSO, SCIM, custom roles and audit logs
- SLA and a named engineer
Compare every plan
One price, unlimited seats. 14-day trial, no credit card.
| Plans | Open Source | Pro Starter | Pro Enterprise |
|---|---|---|---|
| Plan | |||
| Price | Free | $39 / month | Custom |
| 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 |
| Your AI engineers, on agents and ML pipelines | |||
| How it runs | You drive it from your coding agent | On its own, around the clock | On its own, around the clock |
| Agents: reads every trace, flags wrong or missing evals | |||
| ML pipelines: watches every run, flags failures and drift | |||
| Drafts the fix: a new check, eval or pipeline change | |||
| Weekly review in Slack | |||
| Agents and evals | |||
| Trace import | |||
| Evaluators, cohorts and replay | |||
| Scheduled eval runs | |||
| Alerts when an eval starts drifting | |||
| Hosted dashboard | |||
| Pipelines and models | |||
| Pipeline orchestration | |||
| Artifact versioning and lineage | |||
| Webhooks and event triggers | |||
| Run pipelines from the dashboard ↗ | |||
| Managed server, backups and upgrades | |||
| Security and deployment | |||
| SOC 2 Type II and ISO 27001 | |||
| On-prem or air-gapped | |||
| Your own LLM endpoint | |||
| SSO (SAML / OIDC) and SCIM | |||
| Custom roles ↗ | |||
| Audit logs | |||
| Support | |||
| Support level | Community | Dedicated + SLA | |
| Workshops and architecture reviews | |||
| Get started | Sign up free | Talk to us | |
Your agents and your ML, one team of AI engineers
ML pipelines and models
Training, fine-tuning and batch inference on any orchestrator. Your AI engineers watch every run and flag failures and drift.
- Pipeline DAGs, artifact store, model registry
- Version every step, every dataset, every model.
- Run on Kubernetes, Vertex, SageMaker, AzureML
- One pipeline, any orchestrator, no rewrites.
- Fine-tune on your own traces
- A cheaper open model your agent can call. The weights are yours.
AI agents and evals
Your AI engineers find the evals you got wrong or never wrote, then replay your agent to test the fix.
- Insights from your traces
- Finds wrong and missing evals, then turns your answers into checks.
- $39 a month on Starter
- One AI agent and one ML project, one price.
- Seats for your domain experts
- The people who know what good looks like are not a line item.
Starter covers one AI agent and one ML project. Enterprise covers the rest.
Pricing, answered
What do I get for $39?
One AI agent with your AI engineers on it, one ML project with up to 100 pipeline runs a month, and as many seats as you want. It starts with a 14-day trial, no credit card.
What counts as an agent?
An AI agent is anything you want your AI engineers to watch: its code, in one repo or several, and the traces it sends. A LangGraph app with three sub-agents that ships as one product is one agent.
Is there a limit on sessions?
No. Your AI engineers read every session your agent sends. Replays run with your own model keys, so those model costs stay on your bill.
What if we need more agents, projects or runs?
Talk to us. Enterprise covers more agents, more ML projects and more pipeline runs, priced on what you actually run, along with on-prem and governance if you need them.
Why not per seat?
Because everyone who knows what good looks like should be in it: engineers, product, and the domain experts who review sessions. A team of fifteen pays the same as a team of two.
What happens after the 14-day trial?
The trial is full access and needs no credit card. When it ends, add a card to keep going. Nothing in your workspace changes except the billing. Open source stays free, forever.
Do you see my traces?
On Starter, your AI engineers run in our cloud and read traces from your tracing tool through keys you can revoke. ZenML stores only metadata; your data, artifacts and compute stay in your cloud. Self-hosted or on Enterprise on-prem, nothing leaves your infrastructure.
Can we run it on-prem?
Yes, on Enterprise. Your AI engineers and ZenML run inside your own cloud or on-prem, air-gapped if you need it, with your own LLM endpoint.
How is this different from Langfuse, Braintrust or LangSmith?
They show you what happened. Our AI engineers read those traces next to your evals and show you which evals are wrong or missing, then Kitaru replays your agent to test the fix against real sessions. Kitaru imports from Langfuse, LangSmith, Braintrust, Logfire, MLflow or Phoenix, so your tracing tool stays your system of record.
Is there a startup or academic discount?
Yes. Early-stage companies, universities and research institutions can apply for special pricing. Apply here.
Are you startup or academic?
Apply for a special price to access ZenML Pro features for early-stage companies building ML-powered products, universities, research institutions, and educational use cases.
Your VPC, your data
ZenML stores only metadata, so your data and compute stay in your cloud. Your AI engineers read traces through keys you can revoke, or run inside your cloud on Enterprise.
SOC 2 Type II and ISO 27001
We Take Security Seriously
ZenML holds a SOC 2 Type II report and ISO 27001 certification, independently audited for security, availability and confidentiality.
Start with one agent
and your whole team
Your AI engineers read your traces next to your evals and show you what they miss. 14-day trial, no credit card.