Nynch

io.github.peterod99由 GitHub 帳號 peterod99 發布v3.1.0更新於 Oct 26, 2026

Nynch, the AI CRM for consultants and fractionals: contacts, deals and relationship intelligence.

已驗證Streamable HTTP可網頁執行效率與工作流資料與分析商業與電商

概覽

AI 產生的概覽

Nynch 的託管 MCP 伺服器，Nynch 是給顧問與 fractional 使用的 AI CRM，可讓助理搜尋聯絡人與交易、記錄活動並分析關係情報。

可以做什麼
  • 搜尋與讀取 CRM 資料：聯絡人、交易、銷售管道概況、分組以及下一步最佳行動建議
  • 建立與修改 CRM 資料：記錄互動、推進交易階段、建立任務、筆記、聯絡人、公司與交易，以及軟刪除聯絡人
  • 分析關係：權威分數與關係分數、承諾事項、採購決策圈、訊號、趨勢、高風險關係與簡報
  • 檢視公司層級情報：人脈暖路徑、可透過同事觸及的人，以及已成交的客戶專案成果
  • 查詢 Nynch 與已連結應用中受限的唯讀情境資料來源
  • 協調多個代理：註冊代理、共用情境儲存，以及建立與完成交接任務
適用情境
適合把銷售管道與關係紀錄放在 Nynch 的顧問或 fractional 工作者，讓助理不必切換應用就能彙整客戶、找出停滯交易、草擬追蹤訊息或記錄活動。也適用於圍繞共用 CRM 情境執行多代理工作流程的團隊。
執行需求
遠端 Streamable HTTP 端點 app.nynch.com，不需本機執行環境。需要 Nynch 帳號。Claude 可透過 OAuth 登入連線；其他用戶端需要 API 金鑰，以 Authorization bearer 標頭傳送，並在 Nynch 設定中建立。金鑰權限分為 read、read-write 與 admin。每個金鑰每小時上限 1,000 次請求。
安裝前請注意
API 金鑰是敏感憑證，透過 Authorization 標頭傳送；應選擇最小權限，因為 read-write 可建立、更新與刪除紀錄，admin 還包含代理協作。寫入類工具會變更 CRM 資料，刪除類工具可單筆或批次軟刪除聯絡人。伺服器會讀取來自電子郵件、會議與 LinkedIn 的關係情境，公司情報也會揭露同事的關係資料與客戶專案成果。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Nynch，將其新增到工作區。
  2. 連接 所需的 憑證。它們會被私密儲存，絕不會在公開頁面上展示。
  3. 為需要使用其工具的對話啟用該服務。

Web executable，透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式，或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南，先向你展示發布者、端點和登入方式，經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/io.github.peterod99%2Fnynch/install.md，並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前，先告訴我它的發布者、端點和登入方式，等我確認。不要把 API 金鑰寫進命令或我會提交的檔案；問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。 客戶端會在首次連接時完成 OAuth 登入。

{
  "mcpServers": {
    "nynch": {
      "type": "http",
      "url": "https://app.nynch.com/mcp"
    }
  }
}

README

Nynch MCP Server

Website: www.nynch.com · Developer docs: www.nynch.com/developers · Endpoint: https://app.nynch.com/mcp

The official, hosted Model Context Protocol server for Nynch, the AI CRM for consultants and fractionals who win high-value deals based on relationships.

50 tools across 6 categories: CRM read, CRM write, relationship intelligence, firm intelligence, context query, and multi-agent orchestration.

Endpoint

https://app.nynch.com/mcp

Transport: JSON-RPC 2.0 over Streamable HTTP. The server is stateless (no session to manage) and negotiates the protocol version with your client — it speaks 2025-11-25, 2025-06-18, and 2025-03-26.

The older endpoint https://crcrwtjzqzinfzokianr.supabase.co/functions/v1/mcp-server still answers, but it is deprecated. Use https://app.nynch.com/mcp in new configs.

Quick Start

Claude (custom connector, OAuth, no API key)

In Claude, open Settings → Connectors → Add custom connector and enter https://app.nynch.com/mcp. Click Connect, sign in with your Nynch account and approve access. Revoke access any time in Nynch under Settings → Integrations → AI hosts.

Other MCP clients (API key)

1. Get an API Key

Sign up at www.nynch.com, then go to Settings → Integrations → Claude & AI Assistants and create an API key.

Scopes:

  • read — Search, browse, and retrieve data
  • read-write — Read + create, update, delete records
  • admin — Full access including agent orchestration
2. Configure Your Client
Claude Desktop

Add to claude_desktop_config.json:

json
{  "mcpServers": {    "nynch": {      "url": "https://app.nynch.com/mcp",      "headers": {        "Authorization": "Bearer YOUR_API_KEY"      }    }  }}
Cursor

Add to .cursor/mcp.json:

json
{  "mcpServers": {    "nynch": {      "url": "https://app.nynch.com/mcp",      "headers": {        "Authorization": "Bearer YOUR_API_KEY"      }    }  }}
Python (MCP SDK)
python
from mcp import ClientSessionfrom mcp.client.streamable_http import streamablehttp_client
async with streamablehttp_client(    "https://app.nynch.com/mcp",    headers={"Authorization": "Bearer YOUR_API_KEY"}) as (read, write, _):    async with ClientSession(read, write) as session:        await session.initialize()        result = await session.call_tool("search_contacts", {"query": "Acme Corp"})        print(result)
3. Start Using Tools

Your AI client discovers available tools automatically. Ask it to search contacts, check your pipeline, create tasks, or analyze relationships.

Tool Catalog

CRM Read (11 tools)

ToolDescription
search_contactsSearch contacts by name, company, or tags
get_contact_contextFull contact context: profile, interactions, activities, notes, network
search_dealsSearch deals by title, stage, or company
get_deal_contextFull deal context: stakeholders, learnings, coaching actions
search_deals_by_inactivityFind stalled deals with no recent activity
get_pipeline_summaryPipeline health: stage distribution, at-risk deals, upcoming closes
get_next_best_actionsAI-powered recommendations for what to do next
get_who_listPeople to reach now, including identified site visitors. Never mention a specific page in outreach
list_bucketsList all buckets (lists/segments)
get_bucket_contactsGet contacts in a specific bucket
get_bucket_game_queueGet contacts waiting to be sorted into buckets

CRM Write (12 tools)

ToolDescription
log_interactionRecord a contact interaction (call, email, meeting, note)
update_deal_stageMove a deal to a new pipeline stage
create_taskCreate a task linked to a contact or deal
create_noteAdd a note to a contact (supports markdown)
create_contactCreate a new contact (auto-links to company by domain)
update_contactUpdate fields on an existing contact
add_contact_to_bucketAdd a contact to a bucket
create_dealCreate a new deal in the pipeline
create_companyCreate a new company
delete_contactSoft-delete a contact (recoverable)
bulk_delete_contactsSoft-delete contacts by tag or ID list
update_commitmentMark a commitment as completed or update details

Relationship Intelligence (11 tools)

ToolDescription
get_authority_scoreAuthority score across 6 dimensions
get_commitmentsCommitments with Say/Do tracking and due dates
get_buying_committeeStakeholder map: roles, influence, sentiment
get_contact_signalsJob changes, company changes, buying intent, identified website visits (no page paths)
get_interaction_historyFull interaction timeline with direction and sentiment
get_relationship_score5-factor relationship strength score
export_context_briefingAI-ready context briefing for any contact or deal
get_relationship_trendsIs engagement growing, stable, or declining?
get_at_risk_relationshipsContacts with dropping scores or no recent contact
get_deal_intelligenceDeal health: velocity, stakeholder coverage, risk factors
get_weekly_briefingWeekly summary: interactions, changes, recommendations

Firm Intelligence (3 tools)

ToolDescription
get_firm_warm_pathsWho at your firm knows this person or company, ranked by relationship strength. Metadata only, never email content
find_firm_reachable_peoplePeople you can reach through a teammate, strongest path first, with the best teammate to ask for an intro
get_firm_client_outcomesClient work the firm has won: value, retainer, owner, primary contact, referrer, totals by currency

Agent Orchestration (8 tools)

ToolDescription
register_agentRegister an external agent with capabilities and archetype
list_agentsList registered agents with status
agent_heartbeatConfirm agent liveness
write_contextWrite to the shared agent context store
read_contextRead from context store (supports prefix wildcards)
create_handoffRequest another agent to perform a task
get_handoffsGet pending handoffs for your agent
complete_handoffMark a handoff as completed or failed

Context Query (5 tools)

Discover and read bounded slices of Nynch and connected-app context, metadata-first. Read-only — never accepts raw SQL.

ToolDescription
list_context_sourcesList safe read-only context sources available to you
search_context_sourcesSearch context sources by natural-language keyword (metadata only)
describe_context_sourceDescribe one source: purpose, columns, filters, limits, availability
list_context_columnsList selectable columns and allowed filters for one source
query_contextRead a bounded slice from one source via a constrained query object

MCP Resources

URI TemplateDescription
nynch://contacts/{id}Full contact context: profile, score, interactions, network
nynch://deals/{id}Deal context: details, stakeholders, coaching actions
nynch://pipeline/summaryPipeline health overview

Agent Orchestration

Nynch supports multi-agent workflows where AI agents register, share context, and hand off tasks:

  • Register with a name, capabilities, and archetype (producer, consumer, coordinator, critic)
  • Share state via a key-value context store with namespaces and TTLs
  • Route tasks by agent name or capability matching
  • Track liveness with heartbeats

Rate Limits

  • 1,000 requests per hour per API key

Documentation

About Nynch

Nynch is the AI CRM for consultants and fractionals who win high-value deals based on relationships. It remembers relationship context across email, meetings and LinkedIn, and surfaces it when a conversation is worth having.

License

This repository (documentation and configuration) is released under the MIT License. The Nynch MCP Server itself is a hosted service — no local installation required.

來源：README.md，提交 9238f79

工具

0
該伺服器需要登入，因此 SourceWeft 無法在沒有帳號的情況下讀取其工具，也未收錄。已連接帳號所能看到的內容可能因帳號及其權限而異。

版本歷史

2
  1. v3.1.0最新Oct 26, 2026
  2. v1.0.0Sep 16, 2026

評分

已滿足適用檢查項的 85%

該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具15 分

    請讓伺服器在無憑證時回應 tools/list，或發布可在沙箱中無憑證啟動的軟體包。

  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分