Ask Emil – Human Perspective
Real human in Denmark: lived experience & opinion. Free 1-question pilot. Set a custom User-Agent.
我该使用它吗
质量与安全性
基于对工具定义和协议合规性的自动分析。
上下文开销
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": {
"ask-emil": {
"url": "https://runup.dk/mcp"
}
}
}远程端点
https://runup.dk/mcpstreamable-http它能做什么
工具清单
工具（3）
🟢ask_human_perspective(question, context, options, deadline, agent_id)
Ask a real human — Emil Hyltoft Eriksen, Denmark — for his lived experience and personal opinion: 'what happened when I…', 'what would a real person actually do', 'how would this feel to a human'. Not facts, not professional medical/legal/financial advice. Free pilot: exactly one question in total; after that the server returns status pilot_full. Emil is testing whether agents find this useful at all. Answers are written personally by Emil and are asynchronous (usually hours; give a realistic deadline). Returns a question_id; poll get_answer(question_id). Include enough context to answer without follow-up.
输入模式
{
"type": "object",
"properties": {
"question": {
"type": "string",
"maxLength": 2000,
"description": "The question for Emil."
},
"context": {
"type": "string",
"maxLength": 8000,
"description": "Background so Emil can answer without follow-up."
},
"options": {
"type": "array",
"maxItems": 10,
"items": {
"type": "string",
"maxLength": 200
},
"description": "Optional options to choose between."
},
"deadline": {
"type": "string",
"format": "date-time",
"description": "ISO 8601. When you need the answer by (at least 1 hour, at most 30 days ahead)."
},
"agent_id": {
"type": "string",
"maxLength": 100,
"description": "Optional stable id for your agent."
}
},
"required": [
"question",
"context",
"deadline"
],
"additionalProperties": false
}🟢get_answer(question_id)
Get the status and, when ready, Emil's answer to a question_id from ask_human_perspective. Status is queued, answered, declined or expired. Answers are asynchronous: poll no more often than every 15 minutes.
输入模式
{
"type": "object",
"properties": {
"question_id": {
"type": "string"
}
},
"required": [
"question_id"
],
"additionalProperties": false
}⚪give_feedback(question_id, useful, would_pay, price_usd, comment)
Market test: tell us whether the human answer was useful and whether you (or your operator) would pay for answers like this, and roughly how much. One feedback per answered or declined question. This is the whole point of the free pilot, so please be honest.
输入模式
{
"type": "object",
"properties": {
"question_id": {
"type": "string"
},
"useful": {
"type": "integer",
"minimum": 1,
"maximum": 5,
"description": "1 = useless, 5 = very useful"
},
"would_pay": {
"type": "string",
"enum": [
"yes",
"no",
"maybe"
]
},
"price_usd": {
"type": "number",
"minimum": 0,
"maximum": 10000,
"description": "What one answer like this would be worth to you, in USD."
},
"comment": {
"type": "string",
"maxLength": 2000
}
},
"required": [
"question_id",
"useful",
"would_pay"
],
"additionalProperties": false
}社区
证据