MCP server exposing Impiseo SEO data (GSC + crawl + recommendations) to AI agents via a personal API key.
Impiseo MCP demonstrates solid tool design with consistent naming (verb_noun pattern), comprehensive parameter schemas with type definitions and descriptions, and structured output handling. All 13 tools follow a clear verb-first naming convention (get_*, list_*, run_*). Descriptions are present and actionable (avg ~120 chars), exceeding the 10-char minimum. Input schemas are well-formed with Zod validation. However, output schemas are not explicitly documented in tool definitions, responses are validated via envelope() but the actual field structure is not declared to the LLM. Error handling is basic (fail() returns ok:false with a message) but lacks recovery guidance or error classification. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite all tools being read-only. Parameter descriptions could be more prescriptive about constraints (e.g., 'days: 1-90, default 28' is stated but not in a machine-parseable enum or minmax format).
Get the status of the site crawl, including page counts with metrics and content.
Get detailed information about a specific content idea by ID.
Compute query opportunities: ranking queries with low CTR, missing keywords, or high upside potential.
Get GSC overview data including totals, previous period totals, date series, top queries and pages.
Live-fetch and parse a page from the user's site, extracting title, meta description, headings, word count, and optionally full HTML.
Get queries driving traffic to a specific page.
Output schemas not documented in tool definitions. LLMs cannot plan downstream calls or extract fields without knowing response structure. Responses are validated server-side but not declared to the client.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite all 13 tools being read-only. Annotations help LLMs reason about safety and retry logic.
Inferred effective spec: 2025-06-18+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 66 | 2025-06-18+ | v2 |
Get top pages over a window with offset paging, sorted by clicks. Returns enriched data with title and metadata if available.
Get the authenticated user's profile, including email and active property.
Get top queries over a window with offset paging, sorted by clicks.
Get pages ranking for a specific query with their metrics.
Get AI-generated content recommendations for a query or page, including draft titles and meta descriptions.
List all content ideas from the latest run, with optional filtering and pagination.
Run a PageSpeed Insights audit on a URL, returning Lighthouse scores, field data, and failing audits.
Error handling returns only a message string with no recovery guidance or error classification. LLMs cannot determine if an error is retryable, user-fixable, or fatal.
Numeric parameter constraints (days: 1-90, limit ranges, offset max) are documented in descriptions but not enforced as JSON Schema minmax/enum. LLMs may pass invalid values.
Pagination tools (get_queries, get_pages, list_ideas) accept limit and offset but do not document total count or next_cursor in response. LLMs cannot determine if more results exist.