Recipes

Agentic SEO & CRO recipes

Ready-made workflows for pointing AI at your whole blog. Each recipe is a copy-paste prompt plus the safe Specter scaffold — pull your content to markdown, run the AI pass, preview the diff, push it back. Works the same on Ghost, WordPress, Shopify, and Webflow.

Cleanup sweeps

Start here. One-time passes that lift a whole archive in an afternoon.

Cleanup sweeps

The meta description sweep

A copy-paste recipe to rewrite every meta description on your blog with AI: pull your posts to local markdown with Specter, run one prompt across the archive, preview the diff, and push it back.

⏱ ~45 min for 200 posts ~$0.80 in tokens ⚠ Very low (with dry-run)

Cleanup sweeps

The internal link engine

A recipe for bulk internal linking: let AI read your entire archive at once and propose contextual links per post, then review the diff and push with Specter. The highest-leverage on-page SEO move you can automate.

⏱ ~2 hours for 200 posts (mostly review) ~$4–8 in tokens ⚠ Medium — review the dry-run

Cleanup sweeps

The thin content auditor

A read-only recipe to find low-value posts at scale: let AI score your whole archive against thin-content signals and Search Console data, then output a CSV that tells you what to rewrite, merge, or delete.

⏱ ~1 hour for 500 posts ~$2 in tokens ⚠ None (read-only)

Cleanup sweeps

The schema injector

A recipe for structured data at scale: let AI generate FAQ, Article, and HowTo JSON-LD for every post and store it as frontmatter with Specter, so your whole archive becomes eligible for rich results.

⏱ ~40 min for 200 posts ~$1.50 in tokens ⚠ Low (frontmatter only)

Cleanup sweeps

The date-stale refresher

A recipe to fight content decay: scan your archive for stale year references and dated claims, refresh them and the modified date with AI, then preview and push with Specter so Google sees your content as alive.

⏱ ~30 min for 200 posts ~$1 in tokens ⚠ Low

Cleanup sweeps

The broken-link sweep

A recipe to catalog broken internal and outbound links across your whole blog, then repair internal ones to live slugs and flag dead outbound ones — pull, audit, fix, dry-run, push with Spectersync.

⏱ ~40 min for 300 posts ~$1.50 in tokens ⚠ Low — internal fixes are checked against real slugs; you approve the rest

Optimization loops

Run these on a cadence — weekly or monthly — so the gains compound.

Optimization loops

The headline A/B loop

A recipe for rotating blog headlines on a cadence: generate H1 variants with AI, push them with Specter, and let Search Console click-through data pick the winners over time.

⏱ ~30 min setup, ~5 min/week ~$0.50/week (~$9/year) ⚠ Low if you cap rotation frequency

Optimization loops

The brand voice enforcer

A recipe for archive-wide voice consistency: feed AI your style guide, get a per-post report of every deviation with proposed fixes, accept the good ones, and push with Specter.

⏱ ~1 hour for 200 posts ~$3 in tokens ⚠ Medium — voice is subjective

Optimization loops

GSC-driven rewrites

A recipe for data-driven content rewrites: use Search Console to find pages losing ground, hand each one to AI with the queries it ranks for, and rewrite to strengthen the existing signal — then push with Specter.

⏱ ~30 min per 20 pages ~$0.10 per page ⚠ Medium — these are your traffic pages

Optimization loops

The conversion copy sweep

A recipe to turn a blog into a funnel: audit every post through a conversion lens with AI — one clear CTA, a front-loaded promise, scannable subheads — then accept the fixes and push with Specter.

⏱ ~1 hour for 200 posts ~$3 in tokens ⚠ Medium — always review

Optimization loops

The excerpt and alt-text sweep

A frontmatter-only recipe: write custom excerpts and feature-image alt text across your whole blog with AI, for in-feed click-through and image-search traffic — then push with Specter.

⏱ ~30 min for 200 posts ~$1.50 in tokens ⚠ Very low (frontmatter only)

Optimization loops

The AI answer-engine readiness audit

A two-pass recipe: score every post for how well an AI answer engine can extract it, then rewrite the weakest into answer-first articles with summaries and FAQs — pull, audit, fix, dry-run, push with Spectersync.

⏱ ~90 min for 200 posts ~$3 in tokens ⚠ Low — the audit changes nothing; you approve every fix

Optimization loops

The keyword cannibalization report

A read-only recipe that maps which of your posts compete for the same query, then a decision per cluster — consolidate, differentiate, or redirect — applied safely with Spectersync.

⏱ ~45 min for 300 posts ~$2 in tokens ⚠ Low to audit; medium when you consolidate (use a redirect map)

Production

Make new things — clusters, translations, programmatic pages — all as drafts you approve.

Pruning

Cut dead weight safely, with a redirect map so nothing 404s.