LLMs · in SEO content engine
Claude for SEO content engine
Drafting in the SEO content engine stack. Long context, so a brief, a call transcript, and a brand-voice doc all fit in one prompt instead of getting summarized down first.
· yesterday
Ahrefs· Keyword and gap researchPerplexity· Source-backed researchClaude· DraftingSurfer SEO· On-page optimization
Where Claude fits in the workflow
- 2Build the brief
Turn the Ahrefs SERP data into a brief before anyone writes a word: the target keyword, the gaps in the top-ranking pages, and the structure the draft needs to hit.
Prompt · Brief from Ahrefs SERP dataYou are building a content brief for {{keyword}}, monthly search volume {{search_volume}}, Keyword Difficulty {{kd_score}}. Here is what the top 10 ranking pages cover, pasted from Ahrefs (title, word count, headings): """ {{paste ahrefs serp overview or content gap export}} """ Produce a brief with: 1. **Target keyword and intent** - one sentence on what the searcher actually wants. 2. **Gaps** - 3 to 5 things the top 10 pages miss or handle poorly, specific enough to write toward. 3. **Outline** - H2s and H3s, each with a one-line note on what it needs to cover and roughly how many words. 4. **Angle** - the one thing this post says that the top 10 don't. Rules: - No em dashes. Use commas, periods, or parentheses. - Every outline item ties back to a gap or the intent, not a generic subtopic. Output the brief as Markdown, ready to hand to a writer or drop straight into a drafting prompt. - 4Draft
Feed Claude the brief, the sourced research, and a brand-voice doc in one prompt. Long context means all three fit without summarizing any of them down first.
Prompt · Draft from brief and sourced researchWrite a post for {{publication_name}}, audience {{audience}}. Voice: {{e.g. direct, specific, mildly skeptical of hype}}. Brief: """ {{paste the outline and brief from the previous step}} """ Sourced research, use these claims and cite the source inline: """ {{paste perplexity answers with sources}} """ Rules: - Follow the outline's H2s and H3s in order. Do not add or drop sections. - Every claim pulled from the research block gets an inline citation, like "(Ahrefs blog, 2026)". - Open with a specific number or fact, not a definition of the topic. - Never use "delve", "in today's landscape", "it's important to note", "unlock the power of". - No em dashes. Use commas, periods, or parentheses. - Target length: {{word_count_target}} words, matching the brief. Output the full draft as Markdown.
Cost in this stack
$20
Of the $49/mo 1 to 2 posts/wk budget
Tool pricing
$20/mo Pro · Sonnet API $2/$10 per M tokens (input/output)
Alternatives to Claude at this step
Other tools in the SEO content engine stack
Other stacks using Claude
Full-stack with Lovable
Role: Product + scope decisions
Cloud IDE with Replit
Role: Scope + tradeoff calls
UI-first with v0
Role: Product + DB design
AI newsletter
Role: Synthesis + voice
AI newsletter (Substack)
Role: Synthesis + voice
AI thumbnails + ad creative
Role: Prompt engineer
See the full SEO content engine stack
Workflow, costs at three usage tiers, prompts, pitfalls.
Spotted something off?
Wrong price, dead link, stale tool — anything. We review every fix.