DaData MCP server
DaData MCP
DaData official MCP: an agent looks up companies by tax ID, standardizes addresses and identifies a domain or email owner
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- DaData keys draw on a paid request quota, a leaked key can lead to unexpected charges
- The server only reads reference data, it does not change anything in the user's own systems
Install
Manual install
{
"mcpServers": {
"dadata": {
"command": "npx",
"args": ["-y", "supergateway", "--streamableHttp", "https://mcp.dadata.ru/mcp", "--oauth2Bearer", "${API_KEY}:${SECRET_KEY}"]
}
}
}Official connection example from dadata.ru/mcp, the key and secret key come from the DaData personal account.
This is third-party code. Review the repository files before installing.
What it does
The server gives an agent access to DaData's reference data through a remote MCP endpoint. It can look up an organization by tax ID or registration number and get its details, check a counterparty's reliability, standardize and verify an address, identify the owner behind a corporate domain or email, and do geocoding and reverse geocoding. The first company and address lookups are free, after that DaData's paid tiers apply.
Who it is for. For accountants, sales teams and developers who need counterparty checks and address standardization inside a conversation with an agent.
Good fit when
- You need to quickly check a company by tax ID or registration number before a deal
- You need to standardize a customer's address or verify it is valid
- You need to find out who owns a domain or a corporate email
Not a fit when
- You need hundreds of free company checks a month: after the first 50 companies and 100 addresses paid tiers apply
- Your MCP client cannot send a compound Bearer header: the official Claude Desktop example goes through a supergateway wrapper
Example request
Check the company with tax ID 7707083893 and standardize its legal addressLimitations
You need a separate DaData API key and secret key, issued after registration. Only the first 50 companies and first 100 addresses are free, after that companies cost 56000 rubles a year by subscription, domains and emails 7 rubles per request, addresses 0.20 rubles per address. Direct HTTP connection requires sending a compound token in the format key:secret_key, so the official Claude Desktop example uses a supergateway wrapper.
How to disable. Remove the server from your MCP client configuration and revoke or let the DaData keys expire in the personal account.
MCP
- Transport
- http
- Authentication
- API key
| Environment variables | |
|---|---|
| API_KEY required, secret | DaData API key from the personal account |
| SECRET_KEY required, secret | DaData secret key from the personal account |
Security check
- DaData keys draw on a paid request quota, a leaked key can lead to unexpected charges
- The server only reads reference data, it does not change anything in the user's own systems
README in short
The official dadata.ru/mcp page describes the server as a way to connect AI models to company and address data: address https://mcp.dadata.ru/mcp, Bearer-token authorization in the key:secret_key format. It lists the tools: company lookup by tax ID or registration number, address standardization and verification, domain or email owner lookup, counterparty checks and geocoding. It states free limits and prices for companies, domains and addresses, plus connection examples for Claude Desktop via supergateway and for the OpenAI API.
FAQ
How many free company checks do I get?
The first 50 companies and first 100 addresses are free, after that DaData's paid tiers apply.
Why can't I just set a URL and a token like a regular MCP server?
Authorization needs a compound key:secret_key token, so for clients without that support DaData shows an example using the supergateway wrapper.
Related
166 skills for scientific work: bioinformatics, cheminformatics, clinical data, geospatial analysis and 100+ databases
Google's open-source MCP server for databases: ready tools for Postgres, MySQL, BigQuery, Spanner and more, plus custom tools in tools.yaml
Official Hugging Face skills: Hub operations via the hf CLI, datasets, model training, Spaces, evals and deployment
A visualization language for agents and an MCP server: neat charts from a simple semantic spec