Keyphrases Mcp
Pull BERT-based keyphrases from documents in a folder so your agent can draft SEO titles, cluster topics, and tag content without manual keyword guessing.
Overview
keyphrases-mcp is an MCP server for the Launch phase that extracts BERT-based keyphrases from text in a user-specified documents directory.
What is this MCP server?
- stdio MCP server installable via uvx keyphrases-mcp-server (PyPI keyphrases-mcp 0.0.4)
- BERT model-backed keyphrase extraction from text in allowed directories
- Required --allowed-dir filepath scoping which documents the server may read
- Local document batch processing for agent-driven SEO and research prep
- Open-source implementation at github.com/IvanRublev/keyphrases-mcp
- PyPI package version 0.0.4
- Transport: stdio
- 1 required runtime argument: --allowed-dir (filepath)
What problem does it solve?
You have drafts or research files but waste time manually guessing which phrases matter for SEO and positioning.
Who is it for?
Content-heavy solo builders who want on-machine phrase extraction inside MCP workflows before publishing or refreshing pages.
Skip if: Teams needing live SERP volumes, multilingual enterprise SEO suites, or cloud keyword APIs without local Python.
What do I get? / Deliverables
After you configure --allowed-dir and register the stdio server, your agent can return structured keyphrases from those documents for titles, tags, and content plans.
- Keyphrase lists derived from allowed-directory text via BERT pipeline
- Agent-callable local NLP step for SEO and content tagging
- Version-pinned PyPI MCP integration at 0.0.4
Recommended MCP Servers
Journey fit
Keyphrase extraction feeds findability—titles, meta, and content structure—so the canonical shelf is Launch even though you can reuse it during Grow content refreshes. Outputs are keyword and phrase candidates for pages and posts, which maps directly to SEO subphase rather than backend or agent-tooling plumbing.
How it compares
Local BERT keyphrase MCP tool, not a rank tracker or hosted marketing analytics platform.
Common Questions / FAQ
Who is keyphrases-mcp for?
Builders using MCP agents who need quick, model-based keyphrases from local text files for SEO and content structuring.
When should I use keyphrases-mcp?
Use it in Launch (SEO) when drafting or optimizing pages and posts, or when clustering research notes into publishable themes.
How do I add keyphrases-mcp to my agent?
Register stdio with runtime hint uvx, command keyphrases-mcp-server, and required argument --allowed-dir set to your documents directory path.