pinterest-mcp

pinterest-mcp

A guarded remote MCP bridge for publishing original Pinterest Pins, with dry-run safety, pin preview, board listing, and bounded batch publishing.

Category
访问服务器

README

pinterest-mcp

A guarded remote MCP bridge for publishing original Pinterest Pins without living in the Pinterest UI.

Current status

v0.1 implements the Pinterest API + MCP safety core. It is designed for Amy/LabDojo's own Pinterest business account and defaults to dry-run.

MCP tools

  • pinterest_status — read safety/config state
  • list_boards — read boards
  • get_pin — direct Pin readback
  • preview_pin — validate + duplicate-check, never writes
  • publish_pin — create one image Pin only in live mode
  • publish_batch — sequential bounded batch, maximum 10 Pins
  • create_board — create one board only in live mode

There are deliberately no update or delete tools in v0.1.

Safety contract

  1. PINTEREST_WRITE_MODE=dry-run is the default.
  2. The remote MCP endpoint requires MCP_BRIDGE_TOKEN bearer auth.
  3. Pin destinations are restricted to PINTEREST_ALLOWED_LINK_HOSTS (Etsy by default).
  4. Pin images must use HTTPS.
  5. Pinterest field limits are enforced before API calls: title 100, description 800, alt text 500, link 2048 characters.
  6. Before creation, the bridge checks the target board for an existing Pin with the same normalized title + destination link.
  7. After creation, the bridge reads the returned Pin ID back and verifies board, title, and destination.
  8. Batch writes are sequential and bounded to 10.
  9. No secret belongs in Git, Notion, Figma, Drive docs, or chat transcripts.

Authentication

The MVP uses Pinterest Client Credentials because this bridge automates the app owner's own account. Store only the app ID and app secret in deployment environment variables. Pinterest requires two-factor authentication for this grant type.

Requested scopes:

  • boards:read
  • boards:write
  • pins:read
  • pins:write

The server mints and caches access tokens as needed; no long-lived access token is committed.

Pinterest developer prerequisites

  1. Use a Pinterest business account and verify its email.
  2. Enable 2FA for the account/app owner.
  3. Register a Pinterest developer app and request API access.
  4. Trial access can exercise Pin creation but created Pins are visible only to the creator. Standard access is the production target for public automation.
  5. Copy the App ID + App Secret into deployment secrets only.

Environment

Copy .env.example and set:

PINTEREST_APP_ID=...
PINTEREST_APP_SECRET=...
PINTEREST_WRITE_MODE=dry-run
PINTEREST_DEFAULT_BOARD_ID=...
PINTEREST_ALLOWED_LINK_HOSTS=etsy.com,www.etsy.com
MCP_BRIDGE_TOKEN=<random 24+ char secret>

Deploy

The repo is Vercel-ready.

  • health: /health
  • MCP: /mcp

The MCP endpoint uses stateless Streamable HTTP so it works cleanly on serverless infrastructure.

Image delivery

Pinterest can create image Pins from a public HTTPS image_url. Private Google Drive files should remain private and are not valid direct Pinterest media sources.

Planned Remy workflow:

Drive approved master -> posting derivative -> controlled public asset URL -> preview_pin -> publish_pin/publish_batch -> Pinterest readback -> Notion status

The public delivery layer is intentionally separate from source truth. Drive remains the authoritative archive.

Remy Relics

examples/remy-relics-parked-evidence-batch.json contains the four evidence Pins that remain parked after the first four object Pins were manually posted. Their image URLs are intentionally blank until the controlled public media layer is connected.

Local checks

npm install
npm run check

Production rule

Do not flip PINTEREST_WRITE_MODE=live until all of the following are true:

  • Pinterest app access is approved
  • board ID is verified from API readback
  • media URL is the approved final derivative
  • preview_pin shows no collision
  • title/description/alt/link match canon

Then write narrowly, verify the Pin ID, and park the bridge back in dry-run when doing one-off maintenance.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选