Plans and Pricing

Simple, transparent pricing for every team size. Start free and scale with confidence.

Hobby

$0/mo

For developers evaluating and building personal projects.

Free, forever

Compute

Scale-to-zero (cold starts)
Always-on instance

Requests

500K requests/mo
30 req/min rate limit
Burst: 10 req/s for 30s
Overage (hard cut at limit)

Platform

GitHub deploy
NitroStudio included
Usage dashboard
Community support
Get Started Free
Most Popular

Pro

$29/org/mo

For production teams running real MCP workloads.

Billed monthly · $5 per 1M extra requests

Compute

AutoScaling on Spikes
No cold starts
Premium compute nodes

Requests

5M requests/mo included
300 req/min rate limit
Burst: 50 req/s for 60s
Overage requests (10M extra)*

Platform

GitHub deploy
Custom domains
7-day log retention
Usage analytics
Email support
Upgrade to Pro

Enterprise

Custom

Dedicated infrastructure, custom SLAs, and advanced controls.

Starts at $199/mo · volume discounts

Compute

Dedicated node(s)
Always-on, custom min instances
Max instances 10+ instances

Requests & SLA

Custom request quota
Custom rate limits
99.9% uptime SLA
Volume overage discounts

Platform

RBAC & team management
Audit logs
Custom log retention
Dedicated Slack + SLA
Dedicated onboarding
Talk to Sales

* Overage allowed when usage based billing is enabled.

Compare all features

Every detail, side by side.

FeatureHobbyProEnterprise
Instance Configuration
Always-on instance (min instances)✓ min instances=1Custom
Cold startsYes (~3-6s)NoneNone
Scaling
AutoScaling on SpikesUp to 3 instances (on demand)Custom (10+)
Scale-to-zeroFree tier instances only

Overage pricing details

Clear rules for traffic beyond included quotas.

Requests

$5

per 1M requests

Applies to Pro plans when usage-based billing is enabled.

0-10M extra$5 / 1M
10-50M extra$4 / 1M
50M+ extra$3 / 1M

Burst Capacity

Included

on Pro plans

Burst scaling up to 50 req/s for 60s is included in Pro.

Hobby10 req/s, 30s
Pro50 req/s, 60s
EnterpriseCustom

Enterprise Volume

Custom

contract pricing

High-volume teams receive custom overage and committed use discounts.

Committed useAvailable
Long-term contractsDiscounted
Dedicated infraIncluded

Frequently asked questions

Everything you need to know before you deploy.

A request is any single HTTP call to your MCP server, including tool invocations and completion calls. Requests are counted at ingress, so retries count as separate requests. Requests returning HTTP 429 are not counted against quota.

Deploy your first MCP server in minutes.

No credit card required for Hobby. Upgrade when you are ready or talk to us about Enterprise.