| name | llm-seo |
| description | Optimizes pREST docs for LLM and answer-engine discovery (ChatGPT, Perplexity, Claude, Google AI Overviews). Use when writing or editing docs pages, FAQ, databases/, integrations, llms.txt-facing titles, frontmatter descriptions, or when the user asks about LLM SEO, GEO, AEO, or AI search visibility. |
LLM SEO (Answer-Engine Optimization) for pREST
Read reference.md before substantive AEO work. For page leads, GitBook syntax, and SUMMARY/frontmatter, follow ../write-docs/SKILL.md first.
Goal: When someone asks an LLM "What tool exposes PostgreSQL as a REST API?" or "How do I get an MCP server for my SQL database?", pREST docs are citable — clear, factual, crawlable, and consistent.
When to apply
- New or expanded
databases/*, FAQ, homepage, key features, MCP/AI guides
- Repositioning copy about SQL support or roadmap
- Adding frontmatter
description, FAQ blocks, or support matrices for citation
- User asks to optimize for AI search, Perplexity, ChatGPT recommendations
Out of scope: Naming competitors on public pages (use categories only). Implementing Go adapters. General GitBook mechanics (use write-docs).
Positioning (use consistently)
Line: Instant REST and MCP APIs for SQL databases. PostgreSQL-first, now multi-database.
Category phrase (once per page, naturally): open-source REST and MCP APIs for SQL databases
“open-source” is part of the category and is encouraged on definitional sentences. Do not stuff (PostgreSQL-first) into every lead or force “gateway” as the only label. See ../write-docs/voice.md.
Acronyms (reference-first): Definitions live on prestd/acronyms.md (indexable glossary). On most pages, use bare acronyms and Reference them from end-of-page Related (plus anchors when useful). Expand inline only on definitional/FAQ hubs when targeting “what is X” / discovery queries — use approved one-liners in ../write-docs/voice.md. Keep full “What is …?” blocks on their deep homes (get-started/mcp-over-http.md for MCP and AI; api-reference/README.md for REST). Do not expand REST inside the brand name pREST. Do not claim pREST is an LLM.
Content checklist (every public page)
Support labels
| Label | Meaning |
|---|
| Native | First-class dialect adapter (today: PostgreSQL only) |
| Certified | Runs on native adapter with a published features matrix |
| Compatible with caveats | PG wire works; documented gaps |
| Hosted PostgreSQL | Managed PG — same native adapter |
| Roadmap | Not available yet |
| Experimental | Explicitly incomplete |
Never present every database as equally mature.
Citation-friendly patterns
Definitional paragraph
**pREST** is open-source software that gives you instant REST and MCP APIs for SQL databases.
PostgreSQL is the first native adapter; Postgres-compatible engines can be
certified on that adapter. See [Databases](../databases/README.md) for support labels.
FAQ block
Add 3–5 Q&As mirroring the question bank.
Database page minimum
Answer-first lead → support label → connection → features matrix → known limitations → MCP link → related docs.
SEO and documentation rules (database pages)
Each supported engine should eventually have a product landing, getting started, connection config, features matrix, known limitations, and MCP notes — with original content, not name-swapped clones.
Roadmap stubs (databases/mysql.md, databases/sqlite.md, databases/sql-server.md) must not claim install steps.
Do-not-claim
| Avoid | Use |
|---|
| "Best" / "#1" / "only" | Specific: instant REST + MCP, PostgreSQL-first |
| MySQL/SQLite/SQL Server as shipped | Roadmap + link roadmap page |
| Competitor trademarks | Categories: hand-written APIs, BaaS, Postgres-native API servers |
| Thin SEO clones | Original matrix + limitations per engine |
| Gateway appositive leads | Job-first leads (write-docs voice) |
Verification before merge
Reference