QuoteFirst

Get dollar quotes from 200+ LLMs before any work runs; accept, execute, and verify quality.

我该使用它吗

质量与安全性

A
描述质量
97%
模式完整度
75%
命名质量
83%
投毒风险
100%
权限匹配度
100%
协议合规性
100%

基于对工具定义和协议合规性的自动分析。

上下文开销

~552token 数(工具定义)
~301 B典型响应大小
对注意力的影响极小(占 128k 上下文窗口的 0.43%)

这是每次将服务器的工具加载到模型上下文窗口时所消耗的大致 token 数。数值越高,可用于其他任务的注意力就越少。

安装

一键安装

将以下内容添加到你的 `claude_desktop_config.json` 文件中:

{
  "mcpServers": {
    "quotefirst": {
      "url": "https://quotefirst.ai/api/mcp"
    }
  }
}

远程端点

https://quotefirst.ai/api/mcpstreamable-http

它能做什么

工具清单

工具(6)

🟢 只读🟡 写入🔴 删除⚪ 未知
🟡quote(description)

Submit a task description and receive dollar quotes from multiple frontier AI models before any work is done. Returns the request id, quotes with prices in cents, and a recommendation.

输入模式

{
  "type": "object",
  "properties": {
    "description": {
      "type": "string",
      "description": "The task to be quoted"
    }
  },
  "required": [
    "description"
  ]
}
🟢accept(request_id, quote_id)

Accept one quote on a request. The chosen model executes the task. Poll get_deliverable for the result.

输入模式

{
  "type": "object",
  "properties": {
    "request_id": {
      "type": "string"
    },
    "quote_id": {
      "type": "string"
    }
  },
  "required": [
    "request_id",
    "quote_id"
  ]
}
🟢get_deliverable(request_id)

Get the status and deliverable of a request. Status: quoting, quoted, executing, completed, failed.

输入模式

{
  "type": "object",
  "properties": {
    "request_id": {
      "type": "string"
    }
  },
  "required": [
    "request_id"
  ]
}
⚪rerun(request_id)

Quality guarantee: re-run a completed request once for free if its deliverable scored below 70/100 on the automated quality check. Produces a fresh deliverable with the same accepted quote.

输入模式

{
  "type": "object",
  "properties": {
    "request_id": {
      "type": "string"
    }
  },
  "required": [
    "request_id"
  ]
}
🟢spend

Get your QuoteFirst spend this month in cents, total and broken down by model.

输入模式

{
  "type": "object",
  "properties": {}
}
🟢estimate_cost(description, expected_output_words)

Deterministic cost estimate across all models for a task, without using a quote from your allowance. Optionally pass expected_output_words.

输入模式

{
  "type": "object",
  "properties": {
    "description": {
      "type": "string"
    },
    "expected_output_words": {
      "type": "number"
    }
  },
  "required": [
    "description"
  ]
}

社区

评价此服务器

证据

最近观测

已验证未记录版本6 个工具
已验证未记录版本6 个工具