Today we're sharing some important news: we're shutting down Riza's hosted Code Interpreter API on October 1st, 2025. In anticipation, we're disabling new account signups today. Our self-hosted container will continue to operate indefinitely with any existing offline license key.

The Code Interpreter API was an experiment launched in April of 2024 to support software developers looking for a safe place to run LLM-generated code. We built it to be fast and easy to integrate into agentic applications, and it resonated with a set of users who just needed somewhere to run their agent-generated JavaScript or Python as efficiently as possible.

But after reviewing our metrics over the past few months, it's clear that the number of projects that need incredibly fast LLM-generated code execution isn't growing quickly-enough to justify further investment.

We'll turn off the API (https://api.riza.io) and dashboard (https://dashboard.riza.io) on October 1st, 2025. They'll both continue to function as-is for the next two months, but we suggest migrating your code execution workloads to another provider in the next month to avoid any interruptions.

The timeline for deprecating microVM-based custom runtimes remains unchanged: we'll disable support for these runtimes at the end of this week on August 1st, 2025.

We'll be sending an email to all Riza dashboard account holders later today. If you are a Riza user looking to migrate to an alternative service provider, or if you have additional questions about the shutdown, please get in touch on Discord or send an email to hello@riza.io.

Thanks to everyone who tried out the Code Interpreter API and offered feedback along the way. We're grateful for your support.