Build Evidence-Based SEO Audits Without Manual Routing

Codex SEO equips OpenAI Codex with an orchestrator skill that handles natural-language requests like "Do a full SEO check on this website," automatically routing to 26 workflows covering technical SEO, content quality, schema, sitemaps, core web vitals, AI search readiness, GEO, backlinks, local SEO, maps, e-commerce, topic clusters, SXO, hreflang, and SEO drift. It uses 24 specialist agent profiles for targeted execution, shares a cache for evidence reuse across workflows, and prioritizes real data over hallucinations—e.g., skips keyword volume without DataForSEO integration or impressions without Google Search Console. This avoids shallow generic advice ("improve title tags") or scattered multi-tool outputs by producing full audit reports and action plans as structured, deterministic artifacts via local runners, not chat-only responses.

Optional integrations like DataForSEO for research, Google APIs, Firecrawl for crawling, Gemini, and browser-based visual analysis enhance premium checks, but core functionality runs standalone. Slash commands provide agency-grade control, though natural queries suffice for most users.

Install Once, Audit Forever Across Use Cases

Installation takes one command: Mac/Linux uses a single script; Windows has its counterpart. It copies skills into Codex, sets up agent profiles, creates a Python runtime, adds browser support for visuals, runs security/AI checks, and verifies setup—then restart Codex. No dashboard lock-in or black boxes; workflows are readable Markdown files, ideal for learning SEO.

Apply to client audits (agencies), pre-shipment checks (web builders), daily operations (SEOs as a second brain), or post-deployment drift detection ("Find what changed after this deployment"). Examples include "Check this page for schema and core web vitals," "Build an SEO plan for local dental clinic," or "Review this page for AI overviews and ChatGPT search." This Codex-first port of Claude SEO uses distinct agent formats and runtimes but shares the same SEO logic, making it practical for production over chat experiments.