M.Video MCP
Local MCP server for M.Video/Eldorado OmniNet seller API, enabling price/stock updates, product mappings, FBS orders, and shipment management.
README
M.Video MCP
Local MCP server for the M.Video / Eldorado OmniNet seller API.
Больше автоматизаций для селлеров найдешь в моем Telegram-канале: https://t.me/+yNLLxBt6vZ05MDUy
Quick start
This is not a normal desktop app. It is a connector that lets your AI agent work with the M.Video / Eldorado seller API: warehouses, product mappings, prices, stocks, FBS orders, labels, and shipments.
What you need before installation:
- An M.Video / Eldorado seller API key.
- An AI agent or MCP client that supports local MCP servers.
- Node.js LTS installed on the computer where the agent runs.
The easiest way: open your AI agent and send it this message:
Install this M.Video MCP for me: https://github.com/richtargetman/mvideo-mcp
Use Node.js, run npm install and npm run build, then connect dist/server.js as a local MCP server named "mvideo".
Ask me for my MVIDEO_API_KEY only when you are ready to put it into the local MCP environment.
Do not write the real API key into README, Git, screenshots, chat logs, or any public file.
After installation, run mvideo_health and tell me whether the connector is ready.
If you install it yourself, run:
git clone https://github.com/richtargetman/mvideo-mcp.git
cd mvideo-mcp
npm install
npm run build
Then add the MCP client config from the section below and restart your agent.
How to check that everything works:
- Ask the agent:
Check that M.Video MCP is connected. - The agent should call
mvideo_health. - If
apiKeyPresentistrue, the connector sees your API key. - Ask the agent:
Show my M.Video warehouses. - If warehouses are returned, the connector is ready for real work.
Auth
The server reads the seller API key from MVIDEO_API_KEY and sends it as the
api-key HTTP header. Do not store the real key in this repository.
Optional env:
MVIDEO_API_BASE, defaulthttps://omni-net.sellers.mvideo.ruMVIDEO_DEFAULT_LOCATION_ID, default location for stock tools
Tools
mvideo_health- sanitized local configuration; does not call M.Video and never returns the API key.mvideo_merchant_location_list- list seller warehouses.mvideo_product_mapping_list- readoffer_idtoproduct_idmappings.mvideo_price_info- read prices byproduct_idand/oroffer_id.mvideo_price_update- update RUB prices. Supportsdry_run.mvideo_stock_info- read stock byproduct_id,offer_id, and/orlocation_id.mvideo_stock_update- update stock counts. Supportsdry_run.mvideo_order_fbs_list- read FBS orders.mvideo_order_fbs_status_list- read FBS order status updates.mvideo_order_fbs_labels_get- get FBS order labels.mvideo_order_fbs_exemplar_set- set required FBS exemplar/marking data. Defaults to dry-run.mvideo_shipment_create- create an FBS shipment. Defaults to dry-run.
Write tools require an explicit confirmation string for live execution:
mvideo_price_update:confirm="APPLY_MVIDEO_PRICE_UPDATE"mvideo_stock_update:confirm="APPLY_MVIDEO_STOCK_UPDATE"mvideo_order_fbs_exemplar_set:confirm="APPLY_MVIDEO_FBS_EXEMPLAR_SET"mvideo_shipment_create:confirm="APPLY_MVIDEO_SHIPMENT_CREATE"
Use dry_run=true first and only then repeat with dry_run=false and the matching confirm.
How to ask your agent
After the MCP server is connected, you do not need to remember tool names. Write to your agent like you would write to an employee: what to check, what to prepare, and what to change.
Safe read-only examples:
Check if the M.Video connector works.Show my M.Video warehouses.Find product IDs in M.Video for these seller article numbers: SC_BLACK, SC_BLUE.Show current prices for these products.Show current stock balances for these products.Show new FBS orders from M.Video.Get labels for these FBS orders.
Examples that change data:
Prepare a stock update, but do not send it yet: set SC_BLACK to 12 and SC_BLUE to 7.Prepare a price update, but do not send it yet: set SC_BLACK to 499 RUB.Create a shipment for these FBS orders, but first show me the dry-run.
For any price, stock, exemplar, or shipment change, the agent must first show a dry-run plan. If the plan is correct, tell the agent to apply it. The connector will refuse dangerous write actions unless the agent sends the matching confirm string.
Simple rule: checking data is safe; changing prices, stocks, marking data, or shipments must always go through dry-run first.
Agent reference
This section is for AI agents and technical users.
After the MCP server is connected, the agent should inspect available
mvideo_* tools, run read-only checks first, and use live write tools only after
showing a dry-run plan.
Example user requests:
Check that the M.Video MCP is connected and show the active base URL and default warehouse.Show my active M.Video warehouses.Find M.Video product mappings for offer IDs SC_BLACK and SC_BLUE.Show current M.Video prices for offer IDs SC_BLACK and SC_BLUE.Show stock balances for these offer IDs on warehouse 10008903.Prepare a dry-run stock update: set SC_BLACK to 12 and SC_BLUE to 7 on warehouse 10008903.Apply the stock update from the approved dry-run.Prepare a dry-run price update: set SC_BLACK to 499 RUB and old price to 699 RUB.Apply the approved price update.Show new FBS orders and their statuses.Create a shipment for these FBS orders using this idempotency key: <unique-key>. First show dry-run.
Agent usage rules:
- For diagnostics, call
mvideo_health. - For warehouses, call
mvideo_merchant_location_list. - For
offer_idtoproduct_idconversion, callmvideo_product_mapping_list. - For balances, call
mvideo_stock_info. - For prices, call
mvideo_price_info. - For FBS orders, statuses, labels, exemplar data, and shipments, use the matching
mvideo_order_fbs_*ormvideo_shipment_createtools. - Prices in
mvideo_price_infoandmvideo_price_updateare in kopecks, so499 RUBmust be sent as49900. - Stock updates use
MVIDEO_DEFAULT_LOCATION_IDwhenlocation_idis omitted. - Any live write must be preceded by dry-run output and then repeated with
dry_run=falseplus the requiredconfirmvalue.
Run
npm install
npm run build
$env:MVIDEO_API_KEY = "<set locally>"
npm start
MCP client config example
Use your own absolute dist/server.js path after npm run build:
{
"mcpServers": {
"mvideo": {
"command": "C:\\Program Files\\nodejs\\node.exe",
"args": [
"<absolute-path-to>\\mvideo-mcp\\dist\\server.js"
],
"env": {
"MVIDEO_API_KEY": "<set locally>",
"MVIDEO_API_BASE": "https://omni-net.sellers.mvideo.ru",
"MVIDEO_DEFAULT_LOCATION_ID": "10008903"
}
}
}
}
Keep API keys in your local MCP client environment or shell only. Do not commit .env files or credentials.
OpenAPI source: https://omni-net.sellers.mvideo.ru/api-docs.
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。