Cursor ships an Agent SDK with built-in evals
- Source
- x.com
- Date
Cursor released an Agent SDK that treats agents as code. A prompt is a Markdown file and a tool is a TypeScript function, which puts agent behavior in the same repository, review and version history as the rest of a codebase. The SDK ships with evals built in, so each change to a prompt or tool is measured rather than judged by trying it in a chat window. Deployment has two paths: Cursor's hosted infrastructure or a team's own. Cursor's announcement does not go into implementation depth, such as how evals are defined, how hosted runs are billed, or what the self-run path requires, so those details remain open.
The Agent SDK is now available. Define agents as code, with prompts in Markdown and tools in TypeScript. Deploy them on our hosted infrastructure, or run them on your own. Built-in evals let you measure every change and make your agents better over time.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
Checking sign-in…
Loading comments…