This is Hyperinference 

Build your next thing with purpose-built AI inference for coding, from Charm.

Start Free Log In

Hyper, from Charm, is an inference solution with a simple goal: cost-effective, high performance, coding-optimized, open source models, with zero data retention.

Hyper’s great in single player mode, but it's built for teams, and the transition from solo to team use can happen as slow or as quickly as you need it to.

Teams get governance features like master keys and sub-keys, per-model and per-user usage reports, enterprise-grade security, full ownership of generated output, and GDPR compliance.

Pricing

Free
$0 /month

Everyone gets 100 Hypercredits per month.

Get Started
Subscription
$20 /month

250 Hypercredits, refreshing daily.

Get Started
Prepaid Bundles
$5 100 Hypercredits
$10 500 Hypercredits
$20 2,000 Hypercredits

Prepaid Hypercredits never expire.

Get Started

Why hyper?

  • Cost Efficiency Hyper offers LLM inference that's more cost effective than direct provider costs.
  • Specialized for Coding It is specifically designed to streamline agentic AI coding workflows for the Crush developer community, allowing users to complete full workflows that were previously too expensive.
  • An Emphasis on Quality and Reliability Hyper's designed to provide consistent, high-quality AI outputs across its supported coding models.
  • Granular Control Hyper's a pro tool. It offers fine-grained API key management (master and sub keys), giving you and your team better governance over their budgets and usage.

What makes it good?

  • Infrastructure Mastery Unlike a standard proxy or router that simply passes traffic, Hyper is an optimized LLM provider that manages the underlying infrastructure, including orchestration, scheduling, and inference engines.
  • Strategic Partnerships Hyper leverages deep partnerships with AI cloud providers and silicon providers to optimize costs at the hardware and kernel level, rather than just focusing on routing margins.
  • Intelligent Routing Beyond simple proxying, the Hyper roadmap includes intelligent model selection, which automatically delegates tasks to the most cost-effective or specialized model for specific steps in coding workflows.

Hypercredits

Hypercredits are the core of Hyper's usage-based pricing model, where credits map directly to token costs. 1 Hypercredit is currently 5¢, and they're distributed based on your subscription tier.

We believe a prepaid system with credits is easier and more transparent for users to consume versus swiping a credit card and being surprised with a bill later. This is valid both when buying one-off credit bundles or purchasing a subscriptions.