Mcp Email Waterfall Orchestrator

com.mambabuilt由 mambabuilt.com 发布v1.2.1更新于 Nov 7, 2026

Finds and verifies work emails on your own provider keys, with real per-provider spend.

概览

AI 生成的概览

本地 MCP 服务器,使用你自己的查找服务商密钥为联系人查找并验证工作邮箱,并按服务商给出真实花费归因。

可以做什么
  • 用你自己的服务商 API 密钥,按链条依次调用多个查找服务商,为一批联系人查找工作邮箱
  • 每个服务商返回地址后立即验证,把无法投递的“命中”提前标出
  • 按联系人逐行返回结果(含未命中),并附一行汇总:命中的服务商、其在链中的位置、尝试过的服务商、验证结论与消耗单位数
  • 可用运行级服务商单位上限和可选的美元费用上限控制支出;提供自己的费率时以美元报告花费
适用场景
适合已经拥有邮箱查找服务商账号和议定价格、希望让助手对联系人名单批量做补全与验证,并看清各服务商实际花费的场景。也适合需要比较哪家服务商真正带来有效命中的销售或市场数据工作。
运行要求
通过 npx 以 stdio 在本地运行,需要 Node.js。需要 Apify 账号和 APIFY_TOKEN 环境变量,并在工具调用中传入你自己的各查找与验证服务商 API 密钥。需要访问 Apify 及各服务商的网络连接。
安装前请注意
会产生真实费用:Apify 按提交的联系人和找到的地址计费,每次运行还会从你自己的服务商账户扣费。建议设置 maxProviderUnits 和 max_charge_usd 限制成本。服务商密钥在工具调用中传入,应按机密处理。联系人数据(姓名、域名、LinkedIn 链接等)会发送给 Apify 和第三方服务商。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Mcp Email Waterfall Orchestrator,将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Desktop only,通过 STDIO。 STDIO 服务会启动本地进程,因此需要 SourceWeft 桌面宿主。

其他 MCP 客户端

参照 仓库 中的启动说明。

README

Work Email Waterfall MCP Server

[npm] [MCP] [License: MIT]

[Email Waterfall Orchestrator MCP server]

MCP server for the Mamba Labs Work Email Waterfall actor on Apify.

Give it contacts and your own provider API keys. It finds a work email for each, verifies it, and tells you exactly what each provider cost you.

Tool

find_work_email

Returns one flat row per contact, misses included, plus a summary row carrying the run totals: the address found, which provider found it, its position in the chain, every provider attempted, the verification verdict and its source, and the provider units spent.

Bring your own keys

This is the part that makes it different from a credit-based enrichment tool. Mamba Labs supplies no provider keys, never sees your credits, and never resells them. The run spends from your accounts at your negotiated rates, which is why maxProviderUnits exists and why you should set it.

The tradeoff is honest in both directions: you keep your own pricing and your own provider relationships, and you carry your own spend. What you get back is real attribution rather than a credit count, so you can see which provider actually earned its keep.

Give it providerRates and the run reports dollars at your rate. Leave it empty and it reports units only, because your plan tier is not knowable from here and guessing it would produce a confident wrong number.

What happens to a bad key

Nothing dramatic. Every key is validated for free before any paid call. A provider with no usable key is dropped from the chain and reported, rather than failing the run or silently burning attempts. With no keys at all the run still completes, explains which capabilities were unavailable, and charges nothing.

Verification is not an afterthought

Verification runs after each finder that returns an address, not once at the end, because providers do not agree on what "found" means. An address that one provider calls a hit and a verifier calls undeliverable is a miss, and you want to know that before it reaches a sequencer.

find_work_email

InputTypeNotes
contactsobject[]One object per person. Any of full_name, first_name, last_name, company_domain, company_name, linkedin_url, email. A single contact is an array of one.
providerKeysobjectYour own keys, by provider name. Validated for free before any paid call.
providerOrderstring[]Registry entry ids in call order. Empty means the actor's default chain.
verificationOrderstring[]Verifier entry ids in order. Empty means the actor's default.
maxContactsintegerHard ceiling on contacts processed.
maxProviderUnitsintegerRun-level cap on provider credits spent from your accounts. Strongly recommended.
providerRatesobjectMap of provider or provider.endpoint to your cost per unit, so spend is reported in dollars.
concurrencyHintintegerAdvisory. Real pacing comes from each endpoint's own multi-window rate limits.
max_charge_usdnumberOptional cost cap in USD for this run. The actor stops before an item that could take its charges past this amount.

More identifiers means a better chance of a hit. Keyed on a LinkedIn URL the top provider measured 13.7 percent against 9.3 percent on name plus domain, on the same 300 contacts in the same hour.

Setup

json
{  "mcpServers": {    "mamba-email-waterfall-orchestrator": {      "command": "npx",      "args": ["-y", "@mambalabsdev/mcp-email-waterfall-orchestrator"],      "env": { "APIFY_TOKEN": "your-apify-token" }    }  }}

Get a token at console.apify.com/account/integrations. Each call spends from your own provider accounts and consumes Apify credits per contact submitted and per address found, on top of whatever your own provider keys spend.

Pricing

Price per event on the Apify FREE plan, from the live pricing record; paid Apify plans pay less per event.

EventPrice (USD)
Actor Start (apify-actor-start)0.00005
Email Search Submitted (email-search-submitted)0.003
Email Found (email-found)0.02

Also available

This tool is also exposed by the GTM Suite umbrella server, alongside the rest of the Mamba Labs GTM actors, if you would rather run one server than many.

Built by Mamba Labs.

来源:README.md,提交 7bea7b7

工具

1

find_work_email

Find Work Email

Find a work email address for each contact by running a chain of finder providers on YOUR OWN provider API keys, verify each result, and return real per-provider spend attribution instead of an opaque credit count. This is bring your own key: Mamba Labs supplies no provider keys, never sees your credits, and the run spends from your accounts, so set maxProviderUnits. Supply contacts as an array of objects, each carrying any of full_name, first_name, last_name, company_domain, company_name, linkedin_url and email. More identifiers means a better chance of a hit. Supply your keys in providerKeys by provider name. Every key is validated for free before any paid call, and a provider with no usable key is dropped from the chain and reported rather than failing the run; with no keys at all the run still completes, explains what was unavailable and charges nothing. Verification runs after each finder that returns an address, not once at the end. Every contact gets a row including misses, so you can see what was attempted, and a final summary row carries the run totals and spend. Returns flat Clay-ready JSON. Requires an APIFY_TOKEN and consumes Apify credits per contact submitted and per address found, on top of whatever your own provider keys spend.

输入参数contactsmaxContactsproviderKeysproviderOrderproviderRatesmax_charge_usdconcurrencyHintmaxProviderUnitsverificationOrder

版本历史

5
  1. v1.2.1最新Nov 5, 2026
  2. v1.2.0Oct 16, 2026
  3. v1.1.1Oct 6, 2026
  4. v1.1.0Oct 6, 2026
  5. v1.0.2Sep 16, 2026

评分

已满足适用检查项的 90%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通25 分
  • 已声明认证方式10 分

    请在注册中心条目中声明凭据请求头,或保持端点能够响应探测,以便判定其认证方式。

  • 已列出工具必需15 分
  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分