Custom plan

Off-the-shelf plans are great.
Your workload isn't off-the-shelf.

Tell us what you're building. We'll assemble the capacity, models, deployment and support around it — and price it the same day. Starting at $2,000/mo.

Send the brief
1

Send the brief

Two minutes of questions below. No forms-and-forms funnel — the brief goes straight to a solutions engineer.

2

Same-day architecture

We map your workload to models, capacity and deployment (shared, EU-pinned, or VPC) and send a fixed quote within one business day.

3

Pilot week, then ship

A free 7-day pilot on real traffic with an engineer in your Slack. If it doesn't beat your current stack, walk away.

What Custom includes

Everything in Ultra — plus the enterprise layer

⚙

Dedicated capacity

No shared-queue throttling

Reserved GPU pools for your workspace, burst capacity for launches, and latency targets written into the contract.

🔐

VPC & region control

US, EU, or your cloud

Deploy inside your VPC with private endpoints, region pinning, customer-managed keys and full network egress control.

👥

Unlimited seats

SSO / SAML / SCIM

Per-member spend limits, role-based access, shared prompt libraries and consolidated invoicing across departments.

∞

Volume credits

From $0.012 / credit

Pre-purchased credit blocks at volume rates, 12-month rollover, and usage analytics dashboards for finance.

🤝

Solutions engineer

In your Slack

A named engineer who tunes routing, builds custom agents with you and owns your prompt quality — not a ticket queue.

📋

Compliance pack

SOC 2 · GDPR · DPA · HIPAA-ready

Zero-retention agreements with every provider, audit logs with export, and legal review support for procurement.

Brief

Tell us what you're building

Answer what you can — rough numbers are fine. A solutions engineer replies within one business day with an architecture and a fixed quote.

Prefer email? [email protected] for sales, or [email protected] for existing accounts — same humans, same speed.
✓ Brief received. A solutions engineer will reply within one business day.
Enterprise FAQ

Answers procurement asks for

Yes. VPC deployment runs the Sandbox control plane inside your AWS/GCP account with private endpoints to model providers. Prompts and artifacts stay in your VPC; only metered usage metadata leaves it.
Standard DPA at checkout, custom terms on request. SOC 2 Type II report, pen-test summaries and HIPAA-ready configurations available under NDA. We've passed vendor reviews for fintech and health-tech clients.
Your credit rates are locked for the contract term (12–36 months). If provider costs drop, we pass savings through at renewal. If a model you rely on gets deprecated, we migrate it to the successor at no change in credit cost.
That's the default: a free 7-day pilot on real traffic, then a month-to-month pilot contract, then annual terms if you're happy. No pilot-to-production migration work — the environment you test is the environment you keep.