Focus on shipping. We'll meet you at your scale.
Start on Pro with your first AI feature. Move up when the runs do — same canvas, more headroom.
Pro
Your first AI feature, shipping this quarter.
per workspace
- 1 users · 2 projects · 25 pipelines
- 50k runs/mo +
- 30-day trace retention
- BYOK or managed
- Email support
Studio
Multiple features in production, paying customers.
per workspace
- 3 users · 5 projects · 100 pipelines
- 150k runs/mo +
- 30-day trace retention
- BYOK or managed
- Priority email support
Team
Serious volume, multiple eng teams shipping.
per workspace
- 10 users · 10 projects · unlimited pipelines
- 500k runs/mo +
- 90-day trace retention
- BYOK or managed
- Priority support
Enterprise
Regulated workload, dedicated everything.
- Unlimited everything
- Self-hosted or BYOC deployment
- SAML SSO · full audit trail · SLA
- Dedicated Slack + named support
- Custom MSA, security review, DPA
- Platform fee included in contract
BYOK is free. We don’t profit on your tokens.
Choose your inference billing. No markup on tokens, either way.
Bring your own key
Add your OpenRouter API key. We never see or touch your inference dollars. Use any of OpenRouter’s 400+ models.
Managed billing
Top up Noukai credits, we handle OpenRouter for you. One invoice, one dashboard.
- 5% platform fee on top-ups
- No markup on inference at Pro, Studio, Team
Enterprise platform fee is set in your MSA.
What counts as a pipeline run? One trigger, one tick of the meter.
A pipeline run is one execution of one pipeline. The meter ticks once when you trigger it — and only once, however the run plays out.
Sub-pipelines
Count as part of the parent run, not separate runs.
Block retries
Roll into the same run. The meter doesn’t tick again for retries.
Eval runs
Don’t count against your monthly cap. You get 10× more, dedicated to testing.
A “block” is any node in your pipeline — LLM call, tool, memory operation, or control-flow step. The blocks-per-run cap is a fair-use protection against abuse, not a gate on normal usage.
Things you’ll want to know before you commit.
Why “Powered by OpenRouter”?
Because we’d rather build on the best model gateway in the world than reinvent it. 400+ models, automatic failover, OpenRouter’s uptime — all without us taking margin on your tokens.
Can I switch between BYOK and managed billing?
Yes. One click in settings. Your pipelines, traces, and configurations don’t change.
What about data privacy?
Traces and memory are encrypted at rest. Enterprise gets BYOC (Bring Your Own Cloud) deployment so your data never leaves your infrastructure.
Is there a trial?
14 days on Pro, Studio, and Team. No card required. Connect OpenRouter and start building.
How does the annual discount work?
Pay annually and get 2 months free — 17% off the monthly rate.
Can I self-host?
BYOC and self-hosted deployments are Enterprise-only.
What’s overage and when does it kick in?
When you go past your monthly run cap, overage kicks in: your pipelines keep running, and the extras roll onto your next invoice at $5 per 10k extra runs. Same rate on Pro, Studio, and Team.
What happens when I run out of credits or hit my run cap?
We block new pipeline runs and notify you. In-flight runs always finish. Add credits or upgrade to resume.