DwiziDwizi

Start free, scale with control.

Build and run AI tools without infrastructure. Upgrade when you need team governance, higher limits, or private execution environments.

Hobby

Perfect for side projects and experimenting with AI tools and MCP.

$0/month
Start building
  • Unlimited public tools
  • Fair-use limits on shared infrastructure
  • Streaming execution
  • MCP + HTTP endpoints
  • Community support

Runs on shared infrastructure with evolving limits. Tools are expected to be short-lived, stateless by default, and safe for multi-tenant execution.

Early Access

Pro

For professionals shipping production tools with governance, higher limits, and usage controls.

Join the waitlistfor early access
Join Waitlist
  • Everything in Hobby
  • Private tools & team workspaces
  • Higher limits and priority capacity
  • Token controls (scopes, rotation)
  • Usage visibility and budgets
  • Client usage tracking (for chargeback / pass-through)
  • Priority support

Pro is designed for production usage on shared cloud with stronger governance, reporting, and controls.

Enterprise

Dedicated infrastructure for regulated environments. Run Dwizi in your own VPC or on-premise with private networking, governance controls, SSO, and SLAs.

Private deployment (VPC / on-prem)
SAML SSO & SCIM
Audit logs & governance
Dedicated support & SLAs

Frequently Asked Questions

What counts as a “run”?

A run is a single execution of a tool, whether triggered via MCP or HTTP. If a request triggers execution, it counts as a run whether it succeeds or fails.

Do you store tool inputs or outputs?

By default, Dwizi retains audit events only (timestamps, status, and execution metadata). Output and log retention is configurable and can be disabled for compliance-sensitive workloads. Dwizi does not train models on your code, inputs, or outputs.

Is my code secure?

Each run executes in an isolated, ephemeral environment with strict limits and automatic teardown. Dwizi does not train models on your code, inputs, or outputs.

Can I charge my clients based on usage?

Pro introduces client-scoped tokens and per-client usage reporting so you can implement chargeback or pass-through billing models without building your own metering stack.

Does Dwizi only work with MCP?

MCP is a first-class interface, but every tool is also callable over HTTP. The management REST API is available at `https://api.dwizi.com/api` for custom integrations and workflows.

What if I need private networking or compliance?

Enterprise deployments support private networking and governance controls, including SSO/SCIM and audit trails. Output and log retention can be disabled for compliance requirements.