# Ask Jev [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/pollinations/pollinations  
**GitHub Stars:** 5215  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/ask-jev

## Description
Evaluate state with typed choice, score, and probability questions.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What Ask Jev MCP server does

Ask Jev MCP server is intended for structured decision evaluation. Instead of asking for a general narrative answer, a caller provides a question in a supported typed form and receives an evaluation based on that form. The available question categories described for the project are choice, score, and probability.

Choice questions are suited to selecting among specified alternatives. Score questions request a numerical assessment, while probability questions request an estimated likelihood. This makes the project relevant to agent workflows that need a result they can compare, store, or use in a later decision step rather than prose alone.

The supplied repository material places this capability within Pollinations, an open-source AI platform that provides generation APIs and models. It also identifies decision models that score fixed choices and return probabilities. The excerpt does not establish whether every model, endpoint, or Pollinations feature is exposed through the MCP interface, so integration details should be confirmed in the project source.

## How it works

A client submits a typed decision request containing the question and the information needed for its category. The service evaluates that request and returns the corresponding decision-oriented result. The documented behavior distinguishes this flow from ordinary text generation: the goal is a score, probability, or selected option rather than an unrestricted response.

The Ask Jev MCP server is therefore most applicable when the input can be expressed as a fixed decision problem. Examples include comparing named alternatives, assigning a score to a stated situation, or asking for probabilities associated with an outcome. The available material does not describe request schemas, response fields, model selection, authentication, or error handling.

Pollinations' README mentions an `/alpha/decisions` capability for decision requests and lists Perplexity Decider V1.1 27B and Microsoft-Decision-1 as models for typed decisions. That information supports the broader decision-evaluation context, but it does not by itself document the exact MCP server implementation or prove that these models are mandatory for Ask Jev.

## Limitations and notes

The provided documentation is limited. It does not specify a package name, command, configuration file, environment variable, transport, installation procedure, supported MCP clients, or license identifier. It also does not define the exact syntax for choice, score, or probability requests.

Before adopting the Ask Jev MCP server, inspect the repository's relevant package and source files for the current tool names, input schemas, output format, credentials, and startup command. Confirm whether the server calls Pollinations directly, which decision models it can use, and whether returned scores or probabilities include supporting metadata.

Ask Jev MCP server should be evaluated as a typed-decision interface, not as a general-purpose conversational or content-generation connector. Use it when structured evaluation is the requirement; choose another integration when the workflow needs capabilities that the supplied material does not document.

_Full upstream README: https://allmcps.com/mcp/ask-jev/readme_