Hosted MCP server for evidence-backed on-page SEO audits, entity-gap analysis, competitor coverage checks, internal-link opportunities, and page-experience benchmarks.
On-Page.ai SEO MCP is a hosted Model Context Protocol server for search engine optimization work. It gives Codex, Claude Code, ChatGPT, and other AI agents access to structured SEO data for keyword research, competitor analysis, entity gaps, on-page audits, and search-result recommendations through one MCP server.
on-page-mcp.mp4
https://api.on-page.ai/mcp
- Transport: Streamable HTTP
- Authentication: OAuth where supported, API key bearer token for manual clients
- Required OAuth scope:
mcp:seo - Public docs: https://api.on-page.ai/mcp/docs
- Install page: https://api.on-page.ai/install
On-Page.ai gives AI agents structured SEO evidence before they recommend edits. Instead of generic SEO advice, the agent can scan the live URL and compare it against the current search-result cohort for a keyword.
Core workflows:
- Find missing entities and related terms
- Compare competitor topic coverage
- Generate internal-link candidates
- Benchmark page experience against top ranking competitors
- Classify page or text topical focus
- Return customer-safe structured reports for agent reasoning
Use On-Page.ai SEO MCP as part of SEO tools and SEO workflows for AI agents. It helps agents review Google search results, compare SERP competitors, find entity gaps, improve organic visibility, and turn page-level SEO data into specific content recommendations.
The hosted MCP server works with Codex, Claude Code, ChatGPT, and other clients that support MCP setup, API-key authentication, or remote server integrations.
The public API homepage shows the product surface behind the hosted MCP server.
| Tool | Use |
|---|---|
verify_installation |
Zero-credit check that the server is connected, authenticated, and funded |
scan_page |
Default full SEO audit for URL + keyword recommendations |
scan_page_lite |
Faster entity and competitor-cohort scan |
scan_page_deep |
Deeper competitor analysis and optional page-experience benchmark |
classify_text |
Categorize a URL or text into topical buckets |
check_job |
Check async job status |
wait_for_job |
Wait for async job completion |
get_job_result |
Fetch a completed job result |
check_credits |
Check available credits and route costs |
Scans can target 27 country-specific Google SERPs with the region parameter,
and the server publishes report schemas and interpretation guides as MCP
resources. Details: docs/tools.md.
Most users should start here:
https://api.on-page.ai/install
Client-specific docs are also included in docs/install.md.
Use On-Page.ai to scan https://example.com/page for "target keyword".
Return the top missing entities, explain why they matter, and suggest minimal
edits to existing sentences.
Run a deep On-Page.ai scan for https://example.com/page with keyword
"target keyword". Compare the recurring competitor gaps and prioritize the
fixes that are most likely to improve relevance.
More examples: examples/prompts.md
MCP directory metadata lives in:
.cursor-plugin/plugin.jsonmcp.jsonserver.jsonmetadata/directory-submission.jsondocs/directory-listing.md
Cursor plugin assets also include:
This repo should only contain public-facing material. If you are contributing,
read SECURITY.md and do not include:
- private source code
- environment files
- deployment notes
- credentials or tokens
- customer data
- internal admin URLs
- internal review packages
- generated debug output
- Docs: https://api.on-page.ai/docs
- MCP docs: https://api.on-page.ai/mcp/docs
- Changes:
CHANGELOG.md - Contact: team@on-page.ai