Jev Agent
Judgment tools for agents — yes/no, choice and scoring with probabilities, over MCP
About
Jev Agent puts Jev, TypeSafe AI's decision model, behind standard MCP tools. Your agent hands it a piece of text and a question; Jev answers with a probability rather than a paragraph, so the agent can act on a clear answer and stop to ask a person on an unclear one. Seven tools: judge_yes_no, judge_choice, judge_score, compare_versions, plus three essay and resume graders. Each result arrives twice: a one-line summary and the full numbers as JSON. Honesty notes: Jev Agent is independent, not affiliated with TypeSafe, and there is no official TypeSafe MCP server; outputs are probabilistic and require human review. Jev generates no output text, so nothing is billed for output.
Add https://jev-agent.com/api/mcp to your client as a remote Streamable HTTP server and send your Jagent key as Authorization: Bearer <your Jagent key>. Listing the tools needs no key. Exact lines for Claude Code, Cursor and VS Code, and the mcp-remote bridge for clients that only run local servers, are on the Jev MCP page. The same key works for the REST API.
More in MCP Servers
View all
Context7
Up-to-date, version-specific library docs pulled straight into your prompt.
Playwright MCP
Let your AI drive a real browser — navigate, click, fill forms and read pages.
Docker MCP Gateway
Run MCP servers as isolated Docker containers from the signed catalog, behind one gateway endpoint. Needs Docker.