Introducing Cortex: the world's first intelligent blockchain engine

When you’re scaling production apps onchain and performance, reliability, and speed make or break your product, you need high performance infrastructure to support you.

We built something new: introducing Cortex, the first intelligent blockchain engine.

Cortex is a new kind of blockchain engine, designed to deliver what other providers can’t. Trained on trillions of requests and over seven years of data from powering top onchain apps, it provides unmatched performance as you scale.

Powering the new Alchemy developer platform

Cortex runs under the hood across every layer of the our platform. This means every product you use—RPC APIs, Data APIs, Smart Wallets, Rollups—is now powered by Cortex. No setup, no migration, no new tooling required. Just grab an API key and try the brand new developer capabilities powered by Cortex.

Intelligent infrastructure, unmatched performance

Cortex is the result of a complete rethink from first engineering principles. Every layer, from routing and scaling to data consistency and observability, has been redesigned with one goal: to deliver the world’s most intelligent blockchain infrastructure engine.

Let’s break down what Cortex unlocks for you.

Ultra-low latency

Polymarket, the world’s largest prediction market, cut latency from 250ms to 100ms by using our RPC API — powered by Cortex, the intelligent engine behind our developer platform. That drop helped them deliver real-time market data at scale. Here’s how they did it.

This kind of performance isn’t a one-off. It’s the result of deep systems-level engineering in Cortex:

• Pod colocation with Kubernetes affinity ensures latency-sensitive workloads run on the same machine
• Istio locality-aware routing prefers intra-node traffic, cutting out unnecessary network hops
• Microkernel proxy architecture enables modular development without microservice bloat
• Data center-level routing ensures the fastest possible paths for enterprise-grade performance.

That same performance is now available to every developer building on our platform. Our RPC API, powered by Cortex, now delivers sub-50ms average response times, even during peak load or across far-edge regions.

Unlimited elastic throughput

World, powered by Cortex, scales to 30.1M+ users and 598K new accounts per week—while maintaining peak performance. Apps don’t scale linearly, and you shouldn't have to rethink your infrastructure just to handle more users. That’s why Cortex powers our RPC API to support millions of concurrent requests, ensuring your biggest launch days run without throughput bottlenecks.

Here’s how it works under the hood:

• Thousands of globally deployed bare-metal servers purpose-built for high-throughput workloads
• Java Virtual Threads allow each machine to handle thousands of concurrent tasks with low CPU/memory overhead
• AI-powered scaling logic automatically responds to spikes in traffic, memory, or disk usage
• End-to-end observability and tracing enables proactive capacity management and fast troubleshooting

Reliability you don’t have to think about

Uptime isn’t just an infra metric, it protects your users, your brand, and your revenue. Cortex is the reliability engine behind our RPC API, enabling 99.995% uptime with multi-layered failovers, smart recovery logic, and real-time config propagation.

• Regional failovers that reroute traffic instantly when a region degrades
• An isolated emergency backup stack for the rarest edge-case outages
• GitOps-style routing config for safe, controlled rollouts that propagate globally in minutes
• Strict versioning and replication logic to prevent stale reads and maintain data consistency

Robinhood, JPMorgan, Stripe, Coinbase, and Chainlink trust us to power their most critical systems. You won’t need to think about uptime because we already have.

What you can build on top of Cortex

Cortex doesn’t just improve infrastructure—it unlocks entirely new capabilities:

• Unlimited getLogs ranges scans the whole chain for your events 265x faster in a single request
• Block-perfect consistency eliminates jarring errors caused by inconsistent data.
• Smart WebSockets with zero dropped connections, even under heavy load.

That means you can ship simpler code and deliver more powerful features faster.

Ready to build?

Cortex is now powering every part of the our full developer platform.

• Get your API key
• View documentation

The future of onchain development is already here. Let’s build it together.

Frequently asked questions

What is Cortex?

Cortex is our intelligent blockchain engine powering the entire developer platform, including RPC APIs, Data APIs, Smart Wallets, and Rollups. Trained on trillions of requests and over seven years of data, it provides unmatched performance as you scale.

What performance improvements does Cortex deliver?

Cortex delivers sub-50ms average response times, 99.995% uptime, and supports millions of concurrent requests without throughput bottlenecks. It powers apps like World, which scaled to 30.1M+ users and 598K new accounts per week.

How does Cortex achieve ultra-low latency?

Cortex uses pod colocation with Kubernetes affinity, Istio locality-aware routing, microkernel proxy architecture, and data center-level routing to minimize network hops. This enabled Polymarket to cut latency from 250ms to 100ms.

What new capabilities does Cortex unlock for developers?

Cortex enables unlimited getLogs ranges 265x faster in a single request, block-perfect consistency to eliminate data errors, and smart WebSockets with zero dropped connections. This lets you ship simpler code and deliver more powerful features faster.

How does Cortex ensure reliability?

Cortex maintains 99.995% uptime through regional failovers, an isolated emergency backup stack, GitOps-style routing config propagation, and strict versioning logic to prevent stale reads. Companies like Robinhood, JPMorgan, Stripe, Coinbase, and Chainlink trust it for their critical systems.

How does Cortex handle scaling for high-traffic apps?

Cortex uses thousands of globally deployed bare-metal servers, Java Virtual Threads for concurrent tasks, AI-powered scaling logic that automatically responds to traffic spikes, and end-to-end observability for proactive capacity management.

Do I need to migrate or set up anything to use Cortex?

No setup, migration, or new tooling is required. Cortex runs under the hood across every layer of our platform, so just grab an API key to access all the new developer capabilities.

Which products are powered by Cortex?

Every product on our platform is powered by Cortex, including RPC APIs, Data APIs, Smart Wallets, and Rollups.