MCP servers for marketplace search via Apify
Marketplaces MCP Servers
Three MCP servers that search products on Wildberries, Ozon and Yandex Market through Apify actors, with caching and compact output for agent context
Low risk
We rate an entry low when it mostly gives the agent instructions and reference material.
Why this level
- Only searches and reads public product listings, changes nothing
- The Apify token spends the quota of that platform's paid or free plan
Install
Manual install
python wb_mcp_server.pyRuns the Wildberries server; ozon_mcp_server.py and ym_mcp_server.py work the same way for the other platforms.
This is third-party code. Review the repository files before installing.
What it does
Each of the three servers runs a separate Python process with a search tool (search_wb, search_ozon, search_ym) that calls the matching marketplace through the Apify platform and returns priced products. A shared core provides an LRU cache with a TTL and entry limit, exponential-backoff retries on network errors and rate limits, and a compact mode that returns only the top three products to avoid wasting agent context. There are utility tools for checking config and provider health.
Who it is for. For people who want to quickly compare a product's price on Wildberries, Ozon and Yandex Market through an agent without building their own scraper.
Good fit when
- You need a quick price lookup for a specific product on one or more marketplaces
- You need built-in caching and retries to stay within Apify limits
- You need a compact top-3 response instead of the full listing
Not a fit when
- You have no Apify account: the servers do not work at all without an APIFY_TOKEN
- You need seller-side orders, stock or pricing: the servers only search products as a buyer would
Example request
Find an iPhone 17 Pro Max 1TB on Wildberries and Ozon and show the best pricesLimitations
Search does not go directly to the marketplaces but through third-party actors on the Apify platform, so result quality and stability depend on someone else's infrastructure and Apify plan. Concurrency is recommended to stay at one unless your Apify plan allows more. The repository has no license and no requirements file; manual package installation is not documented.
How to disable. Stop the server process and remove it from your MCP client config.
MCP
- Transport
- stdio
- Authentication
- API key
| Environment variables | |
|---|---|
| APIFY_TOKEN required, secret | Access key for the Apify platform, required by all three servers. |
| MAX_CONCURRENT_SEARCHES | Concurrency limit for Apify requests, defaults to 1. |
| SEARCH_CACHE_TTL_SEC | Search result cache TTL in seconds, defaults to 900. |
Security check
- Only searches and reads public product listings, changes nothing
- The Apify token spends the quota of that platform's paid or free plan
README in short
The Russian README describes the shared core for caching, throttling and Apify API calls, gives an environment variable table with APIFY_TOKEN, a concurrency limit and timeouts, lists the search tools and the config/status utility tools, and includes a ready system prompt for the agent.
FAQ
Do I need an Apify token?
Yes, APIFY_TOKEN is required; without it none of the three servers can search products.
What does compact mode do?
Setting compact: true on a search returns only the top three products instead of the full list, saving context.
Related
CLI, skill and Python library for browser control: the agent clicks, fills forms and reads pages over CDP
A skill that gathers the last 30 days of discussion on a topic from Reddit, X, YouTube, HN, Polymarket and GitHub into one brief
The Chrome DevTools team's official MCP server: the agent drives a live Chrome, reads network and console, and records performance traces
A fast Rust CLI for agent browser automation: accessibility snapshots with element refs, an MCP server and skills