# Jagent

> Jagent is a freemium MCP server for ai models and llm tooling. Hosted yes/no, choice and rubric judgments with probabilities, plus essay graders. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Freemium
- **Platforms:** API, Web
- **Website:** https://jev-agent.com/jev-mcp
- **MCP Registry name:** com.jev-agent/jagent
- **Categories:** AI models and LLM tooling, developer tools
- **Listing updated:** 2026-11-17

## About

Jagent wraps Jev, a decision model from TypeSafe AI, in a hosted MCP server so an agent can get a second opinion on its own work. Instead of a paragraph, each tool returns a probability. judge_yes_no reads 0.6 and above as yes, 0.4 and below as no and anything between as too close to call. judge_choice picks among two to fifty options, judge_score places text on a scale of two to ten levels, and compare_versions asks about two drafts in both orders and averages, reporting a tie if the orders disagree. Three further tools grade essays and resumes.

The site suggests guard patterns, such as asking whether a planned command can be undone before running it, or whether test output shows every test passed before claiming success. The tools judge only; the agent's own model still writes.

Listing tools needs no key, while calls need a Jagent key as a bearer token. Credits are one-time packs, one per 1,000 input tokens with output free, and 5 free credits each month. The service says it is independent and not affiliated with TypeSafe AI.

## Features

- **Yes/no judgment:** Returns the probability of yes with a clear cut-off for yes, no and unclear.
- **Multiple choice:** Chooses among up to 50 options with a probability for each.
- **Rubric scoring:** Places text on a 2 to 10 level scale.
- **Version comparison:** Compares two drafts in both orders and reports ties.
- **Graders:** Essay and resume grading tools.
- **Dual output:** A one-line summary plus full JSON numbers.

## Good for

- Check an action is reversible before an agent runs it
- Verify test output really shows success
- Route a task to the best-fitting tool
- Score a draft reply before sending it

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http jagent https://jev-agent.com/api/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://jev-agent.com/api/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "jagent": {
      "url": "https://jev-agent.com/api/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "jagent": {
      "type": "http",
      "url": "https://jev-agent.com/api/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "jagent": {
      "serverUrl": "https://jev-agent.com/api/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "jagent": {
      "type": "http",
      "url": "https://jev-agent.com/api/mcp"
    }
  }
}
```

## Tools

- `judge_yes_no`: Probability of yes for a question about some text.
- `judge_choice`: Choose among 2 to 50 options with probabilities.
- `judge_score`: Place text on a 2 to 10 level scale.
- `compare_versions`: Compare two versions in both orders and pick the stronger.

## FAQ

### How is it billed?

One credit per 1,000 input tokens, minimum one per call, with no charge for output; credits come in one-time packs that never expire.

### Is there a free tier?

Yes, 5 credits a month, spent before purchased credits.

### Does it write or edit text?

No. Every tool returns a judgment; your agent's own model does the writing.

### Do I need a TypeSafe key?

No, a Jagent key is used, and TypeSafe keys cannot authenticate to this server.

### Is Jagent free?

Jagent has a free plan, with paid plans for more usage or features.

### How do I add Jagent to Claude Code?

Run this command in your terminal: claude mcp add --transport http jagent https://jev-agent.com/api/mcp

### What are the best alternatives to Jagent?

The closest alternatives to Jagent are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [GitHub](https://ai-powered.com/mcp/github-2): GitHub is an open-source MCP server for developer tools.
- [Serena MCP: the IDE for your agent](https://ai-powered.com/mcp/serena-mcp-the-ide-for-your-agent): Serena MCP: the IDE for your agent is an open-source MCP server for developer tools.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Ouroboros](https://ai-powered.com/mcp/ouroboros): Ouroboros is an open-source MCP server for developer tools.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Concord MCP](https://ai-powered.com/mcp/concord-mcp): Concord MCP is a freemium MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/jagent
