SynapseSynapse
← Back to Gallery
WorkflowVerified

Blog Post Generation Pipeline

A multi-step workflow that takes a topic and target audience, researches the subject, creates an SEO-optimised outline, drafts the full post with headings and examples, then polishes tone and readability. Includes a meta description and social preview.

896.7k

Pipeline Overview

This multi-step workflow turns a topic + audience pair into a publish-ready blog post in four stages:

  1. Research — web search + source synthesis
  2. Outline — SEO-optimised H2/H3 structure with keyword targets
  3. Draft — full prose with examples, analogies, and data
  4. Polish — tone, readability score, meta description, social previews

Each stage can be run independently or chained end-to-end.

Four-stage pipeline from topic to publish-ready post
Four-stage pipeline from topic to publish-ready post

Stage Prompts

Stage 1 — Research

Topic: {topic} Target audience: {audience} Goal: Identify the 5 most important sub-topics, key statistics, and authoritative sources. Output as structured JSON: { "sub_topics": [...], "key_stats": [...], "sources": [...] }

Stage 2 — Outline

Given the research below, create an SEO-optimised blog post outline. Primary keyword: {primary_kw} Secondary keywords: {secondary_kws} Requirements: - H1 title (include primary keyword) - Introduction hook (question or statistic) - 4–6 H2 sections with H3 sub-points - Conclusion with CTA - Estimated word count per section

Stage 4 — Polish

Review the draft below and: 1. Improve any passive-voice sentences 2. Replace jargon with plain language where possible 3. Add a meta description (150–160 chars, include primary keyword) 4. Write 3 social preview variants (Twitter, LinkedIn, newsletter) 5. Rate readability: Flesch–Kincaid grade level target ≤ 10

Supported Output Formats

  • markdown — clean .md file ready for any CMS
  • notion — Notion API block format
  • wordpress — Gutenberg-compatible JSON blocks
  • html — standalone article with semantic tags

Discussion (0)