概览
What is On-Page.ai SEO MCP?
On-Page.ai SEO MCP is a hosted Model Context Protocol (MCP) server that provides AI agents with structured, evidence-backed search engine optimization data. It enables keyword research, competitor analysis, entity-gap detection, on-page audits, and search-result recommendations. The server is designed for developers and SEO professionals using AI coding assistants like Codex, Claude Code, or ChatGPT.
How to use On-Page.ai SEO MCP?
Install the server from the quick-install page at https://api.on-page.ai/install. Supply authentication via OAuth (with the required scope mcp:seo) or an API key bearer token for manual clients. The server URL is https://api.on-page.ai/mcp and uses Streamable HTTP transport. Invoke tools such as scan_page, scan_page_lite, scan_page_deep, classify_text, check_job, wait_for_job, get_job_result, or check_credits from any MCP-compatible client.
Key features of On-Page.ai SEO MCP
- Scans live URLs and compares them against search-result cohorts.
- Finds missing entities and related terms for target keywords.
- Compares competitor topic coverage and identifies gaps.
- Generates internal-link candidates automatically.
- Benchmarks page experience against top-ranking competitors.
- Provides customer-safe structured reports for agent reasoning.
Use cases of On-Page.ai SEO MCP
- Automate SEO audits and content recommendations inside AI coding assistants.
- Improve organic visibility by identifying entity gaps and competitor coverage.
- Generate targeted content briefs based on real-time search result analysis.
- Benchmark page experience metrics against top SERP competitors.
- Integrate SEO evidence directly into code review and writing workflows.
FAQ from On-Page.ai SEO MCP
How do I authenticate with the server?
Use OAuth with the mcp:seo scope where supported, or provide an API key bearer token for manual clients. The server URL is https://api.on-page.ai/mcp.
What tools are available?
Eight tools are provided: scan_page (full audit), scan_page_lite (faster scan), scan_page_deep (deep analysis), classify_text, check_job, wait_for_job, get_job_result, and check_credits.
What data does the server access?
It accesses structured SEO data by scanning live URLs and comparing them against the current search-result cohort for a specified keyword. It returns entity gaps, competitor coverage, internal-link opportunities, and page-experience benchmarks.
Is there a credit system?
Yes. The check_credits tool shows available credits and per-route costs. Each scan or analysis consumes credits based on the tool used.
Which AI clients are supported?
The server works with Codex, Claude Code, ChatGPT, and any client that supports MCP setup, API-key authentication, or remote server integrations.
