Zopaf MCP

MCPAI代理

提供方 rjandino

AI代理 未标注 无需鉴权 社区 / 第三方

Negotiation math engine for AI agents. Computes Pareto frontiers, generates iso-utility counteroffers, and infers counterpart priorities. 9 tools for any multi-issue negotiation. Zero LLM tokens — pure MILP optimization.

一句话选型结论

按资料字段整理

Zopaf MCP 属于AI代理类 MCP 服务器,传输方式未标注,无需鉴权,没有现成安装命令,需按官方文档接入。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类AI代理 传输方式未标注 鉴权无需鉴权 部署方式支持远程部署
GitHub 星标 未标注 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

AI代理
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

安装与配置

未标注 无需鉴权
上游未给出传输方式
  1. 1
    准备运行环境

    确认你的客户端支持远程端点,并拿到服务地址。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
给代理接外部能力

让代理在任务中途调用搜索、数据库等真实工具。

多步任务编排

把一串操作拆成步骤,由代理按顺序执行。

自动化重复流程

把固定流程交给代理跑,减少手工操作。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 资料标注无需鉴权,装好就能连。

先想清楚的地方

  • 远程端点依赖网络与服务方可用性,注意密钥与调用额度。
  • 没有现成安装命令,需要按官方文档自己填启动命令和参数。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在AI代理场景里让模型直接调用外部能力的人。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Zopaf MCP当前 未标注 未标注 支持远程部署 无 2026-05-17
Filesystem 560c4363 13 未标注 本地部署 无 2026-05-17
Server Everything (everything) 11 stdio · 本地进程 本地部署 有 2026-05-17
Time 7365cb6d 2 SSE · 服务端推送 远程部署 有 2026-05-18
Everything Claude Code 未标注 stdio · 本地进程 本地部署 有 2026-05-17
Openmemory 557e1585 未标注 stdio · 本地进程 本地部署 有 2026-05-17

服务器档案

AI代理
字段来自项目仓库与公开索引,中文说明由本站标注
提供方rjandino上游标注的维护方 rjandino
分类AI代理本站按上游分类展示
连接方式未标注上游未给出传输方式
鉴权方式无需鉴权远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时未标注 · 未标注本地安装时用到的包标识与运行时
部署类型支持远程部署上游给出的部署形态
最近核验2026-05-17上游最后核验时间

详细介绍

6.4 KB
来自项目仓库与上游目录的原文,未做删改

Zopaf Negotiation Engine -- MCP Server

A negotiation math engine exposed as MCP tools that any AI agent can call. Zopaf computes Pareto frontiers, generates iso-utility counteroffers, and infers counterpart priorities from their reactions -- all through pure MILP optimization. Zero LLM tokens burned. The calling agent handles the conversation; Zopaf handles the math.

Quick Start

Claude Desktop

Add to your Claude Desktop configuration (claude_desktop_config.json):```json
{
"mcpServers": {
"zopaf": {
"type": "streamable-http",
"url": "https://zopaf-mcp-production.up.railway.app/mcp"
}
}
}


### Claude Code```bash
claude mcp add zopaf --transport streamable-http https://zopaf-mcp-production.up.railway.app/mcp

Generic MCP Client

Connect to the Streamable HTTP endpoint:```
URL: https://zopaf-mcp-production.up.railway.app/mcp
Transport: Streamable HTTP


## Tools Reference

| Tool | Description |
|------|-------------|
| `create_session` | Create a new negotiation session. Returns a session_id used by all other tools. |
| `add_issue` | Add a negotiable issue/term with options ordered worst to best for the user. |
| `set_issue_range` | Set the acceptable range for a numeric issue, enabling 0-100 scoring. |
| `record_preference` | Record that the user prioritizes some issues over others. Updates the weight model. |
| `set_batna` | Record the user's alternatives if the deal falls through. Determines leverage. |
| `generate_counteroffers` | Generate 3 iso-utility counteroffers to present simultaneously. |
| `process_counterpart_response` | Process the counterpart's reaction to infer their priorities and generate a round-2 offer. |
| `analyze_deal` | Score a specific deal against the Pareto frontier. Shows value captured and suggested trades. |
| `get_negotiation_state` | Get current model state: issues, weights, BATNA, frontier size, and recommended next step. |

## How It Works

1. **Create session** -- Initialize a new negotiation model with `create_session`.

2. **Add issues** -- Define the terms on the table with `add_issue`. Each issue includes options ordered worst to best for the user (e.g., Salary: `['$150K', '$160K', '$170K', '$180K']`).

3. **Set ranges** -- For numeric issues, call `set_issue_range` to map values onto a 0-100 scoring scale.

4. **Record preferences** -- Call `record_preference` as you learn what the user cares about. Each call updates the internal weight model.

5. **Set BATNA** -- Use `set_batna` to record alternatives. The number and quality determines leverage strength and anchoring strategy.

6. **Generate 3 counteroffers** -- Call `generate_counteroffers` to produce three packages that are equally good for the user but structured differently. Present ALL THREE simultaneously. Never lead with one and fall back to another.

7. **Process counterpart response** -- Call `process_counterpart_response` with which package they preferred and what they pushed back on. The engine infers their hidden priorities.

8. **Get round-2 offer** -- Returns a refined offer on the efficient frontier, with value split weighted by leverage.

## Example: Job Offer Negotiation```
create_session
-> {"session_id": "a1b2c3d4"}

add_issue(session_id="a1b2c3d4", issue_name="Salary", options=["$150K", "$160K", "$170K", "$180K"])
add_issue(session_id="a1b2c3d4", issue_name="Equity", options=["0.1%", "0.25%", "0.5%", "0.75%"])
add_issue(session_id="a1b2c3d4", issue_name="Signing Bonus", options=["$0", "$10K", "$20K", "$30K"])
add_issue(session_id="a1b2c3d4", issue_name="Remote Work", options=["On-site", "Hybrid", "Fully Remote"])

set_issue_range(issue_name="Salary", worst_acceptable=150000, best_hoped=180000,
    option_values={"$150K": 150000, "$160K": 160000, "$170K": 170000, "$180K": 180000})
-> {"scores": {"$150K": 0.0, "$160K": 33.3, "$170K": 66.7, "$180K": 100.0}}

record_preference(preferred_issues=["Salary", "Equity"], over_issues=["Signing Bonus", "Remote Work"])
-> {"learned_weights": {"Salary": 0.345, "Equity": 0.345, "Signing Bonus": 0.155, "Remote Work": 0.155}}

set_batna(alternatives=["Competing offer from Company B at $165K", "Stay in current role"])
-> {"leverage_strength": "strong"}

generate_counteroffers(target_satisfaction="ambitious")
-> {
  "counteroffers": [
    {"label": "A", "terms": {"Salary": "$180K", "Equity": "0.25%", "Signing Bonus": "$10K", "Remote Work": "On-site"}},
    {"label": "B", "terms": {"Salary": "$170K", "Equity": "0.5%", "Signing Bonus": "$0", "Remote Work": "Hybrid"}},
    {"label": "C", "terms": {"Salary": "$160K", "Equity": "0.75%", "Signing Bonus": "$20K", "Remote Work": "On-site"}}
  ]
}

process_counterpart_response(preferred_package="B", pushback_issues=["Equity"])
-> {
  "counterpart_priorities_inferred": {"Equity": 0.571, "Salary": 0.143, ...},
  "round_2_offer": {"Salary": "$180K", "Equity": "0.25%", "Signing Bonus": "$20K", "Remote Work": "Hybrid"},
  "value_split": "User gets 75% of surplus"
}

The engine inferred that the counterpart cares most about equity (57% of their weight). The round-2 offer concedes on equity -- where it costs the user less -- and captures value on salary and signing bonus. Both sides improve. The user captures 75% of the surplus based on their strong BATNA.

Use Cases

  • Job offers -- Salary, equity, bonus, title, remote work, start date, PTO
  • VC term sheets -- Valuation, board seats, liquidation preferences, anti-dilution, pro-rata rights
  • Real estate -- Price, closing date, contingencies, repairs, inclusions, rent-back periods
  • Vendor contracts -- Price, SLA guarantees, payment terms, exclusivity, renewal clauses
  • Salary negotiations -- Base pay, bonus structure, review timeline, scope of role
  • Business partnerships -- Revenue split, IP ownership, decision rights, exit clauses, territory
  • Legal settlements -- Monetary terms, non-disclosure terms, admission of liability, timeline

Why Zero Tokens?

Zopaf is a math engine, not a language model. It runs MILP optimization and combinatorial scoring -- operations that are computationally cheap but tedious for an LLM to attempt in-context.

Your agent's LLM handles the conversation with the user, asks the right questions, and explains the strategy. Zopaf handles the optimization -- computing Pareto frontiers, generating iso-utility packages, solving preference weights from revealed choices, and positioning offers on the efficient frontier.

You bring the brain. Zopaf brings the calculator.

常见问题

7 问
答案由本站依据条目字段整理,不引用上游原文
Zopaf MCP 怎么安装?

资料没有给出安装命令,请到官方仓库或文档按说明接入;这类条目多数是生态应用,或尚未发布可安装包。

Zopaf MCP 怎么在客户端里配置?

它的传输方式未标注,在客户端 MCP 配置里按官方文档填写启动命令和参数,保存后重启客户端生效。

Zopaf MCP 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Zopaf MCP 需要 API Key 吗?

资料标注为无需认证,装好就能连;如果实际调用时报鉴权错误,按官方文档补上凭据。

Zopaf MCP 能提供哪些能力?

资料显示能力标签包括 AI代理、工作流自动化、模型集成。具体工具清单以官方文档为准。

Zopaf MCP 需要什么运行环境?

属于远程端点,不需要在本地装运行时。

Zopaf MCP 是本地还是远程?

它是远程端点,连服务地址即可。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 2026-05-17。