DataForSEO MCP
A thin wrapper around DataForSEO's official MCP server that adds a searchable docs index and normalized errors. If you only need the raw API, install dataforseo-mcp-server directly.
Install
Claude Code:
claude mcp add dataforseo -e DATAFORSEO_USERNAME=your-login -e DATAFORSEO_PASSWORD=your-password -- npx -y @scalably-io/dataforseo-mcp
Codex:
codex mcp add dataforseo --env DATAFORSEO_USERNAME=your-login --env DATAFORSEO_PASSWORD=your-password -- npx -y @scalably-io/dataforseo-mcp
Claude Desktop: this wrapper ships on npm only (its upstream dependency makes a desktop bundle over 100 MB). Install DataForSEO's own server from https://github.com/dataforseo/mcp-server-typescript instead; the docs-search tools are the only thing you lose.
Setup
- Create or open a DataForSEO account at dataforseo.com.
- Read the API login (an email address) and API password from the dashboard.
api_request calls the live DataForSEO API and can incur cost per DataForSEO's own pricing; the three docs_* tools are free documentation lookups.
Tools (4)
| Tool | What it does |
|---|
docs_index | Fetch the DataForSEO API documentation index (llms.txt), optionally filtered by section |
docs_list_sections | Return available DataForSEO API documentation section names |
docs_search | Fetch DataForSEO API documentation from a documentation URL |
api_request | Make an authenticated request to the DataForSEO API (method, path or url, body, and an optional noAiMode flag for the full, non AI-optimized response) |
Use docs_list_sections or docs_index to pick an API family, then docs_search for the exact current endpoint path, method, request body and pricing before calling api_request.
Configuration
| Variable | Required | Purpose |
|---|
DATAFORSEO_USERNAME | yes | DataForSEO account login (email) from the dashboard |
DATAFORSEO_PASSWORD | yes | DataForSEO account API password from the dashboard |
NODE_ENV, DATAFORSEO_PROXY_TEST_CHILD | no | Test suite only: with NODE_ENV=test the wrapper spawns the fake child named by DATAFORSEO_PROXY_TEST_CHILD instead of the upstream server. Never set them in normal use |
Reply shape
Every tool returns plain JSON with status (succeeded, partial, no_op), operation, summary, target, result, proof, warnings, recovery. Failures surface as a tool error whose text is <code>: <message> <hint>. api_request normalizes DataForSEO's per-task status codes (pending tasks, no-result tasks, partial failures, and vendor error codes) into that shape instead of returning DataForSEO's raw task envelope unexamined.
Limits
api_request is a real, billable call to the DataForSEO API; DataForSEO's own account balance and rate limits apply. A failed call to a paid endpoint is reported with an ambiguous_paid_request error when the outcome could not be confirmed, since retrying blind could double the charge.
Verify
Each release lists the package version and the production commit it was derived from in CHANGELOG.md; npm publishes with provenance from GitHub Actions. CI runs the tests and a clean install of the packed tarball on every push.
Privacy Policy
This server runs locally, on your machine, under your own credentials. It collects no personal data, contains no telemetry, stores nothing persistently, and talks only to the vendor API it wraps. No third party, including Scalably, receives your data. Contact: [email protected]. Canonical copy: https://scalably.io/connector-privacy.html
License
MIT. Copyright Scalably.
