Jagent

We built Jagent for the moments when a clear score against specific criteria is more useful than another page of AI advice.

You can evaluate an essay draft, compare a resume with a job description, or define your own rubric. The browser tools let you review the criteria before running a check. REST API, MCP, and CSV batch workflows support integrations and repeated evaluations.

The essay grader offers three trials without an account; further usage is credit-based. Results are AI estimates that need human review.

We would love feedback on the rubric editing experience and which evaluation workflows would be most useful to you.