🎯
Strategy Agent
Interviews you (or reads env vars), maps competitors, finds content gaps, stores pillars that guide every future run. Uses GPT-4o for deep reasoning.
Run once · manual or CI
Blogging Agent is a multi-agent Python pipeline that researches SEO topics from Google Search Console & GA4, writes and edits full articles, publishes to your site, monitors performance, and rewrites underperformers—zero manual steps after a one-time setup. We built it for SellOnLLM; you can fork it for any static site or markdown CMS.
Free tier friendly · Groq for writing · GPT-4o for strategy · Gmail run reports optional · Web + CLI run logs
Each agent has a narrow job. Together they replace a content team’s research, writing, editing, and maintenance loop.
Interviews you (or reads env vars), maps competitors, finds content gaps, stores pillars that guide every future run. Uses GPT-4o for deep reasoning.
Pulls GSC queries with high impressions but weak CTR, checks competitor sitemaps, validates keywords via SERP analysis, deduplicates against your archive.
Produces full markdown articles with hooks, H2/H3 structure, internal link suggestions from your post memory index, and fresh web research.
Reviews drafts against the research brief—SEO placement, clarity, engagement—and edits in place with notes on what changed and why.
Reads queue size, ramp limits, and GSC opportunity scores to decide: publish now, fill the queue with research, or wait.
Daily GSC + GA4 snapshots per post. Computes health scores and flags underperformers for the corrector.
Chooses rewrite, retitle, requeue, or wait for flagged posts based on real ranking and traffic data.
Jaccard fingerprint index for internal links and duplicate blocking—no vector DB required. State persists in SQLite across CI runs.
SellOnLLM integrates this agent into our main repo with a markdown→HTML publish bridge, sitemap updates, Gmail run reports, and a live run dashboard. Fork blogging-agent for a standalone setup, or study our integration in vipulawl/sellonllm.