Agents-run SEO pipeline: 8 draft PRs in two weeks, most runs under $1
Pitiful-Surround-285 · reddit · 2026-09-02
A team shares two weeks of running an agentic SEO content pipeline: 8 draft PRs, 1 merged, 1 closed, 6 in review, with most runs under $1 including images.
Setup
- Two agents + one skill file + one prompt + one schedule: a writer agent does keyword research and drafting, briefing a second agent for figures and images; they hand off in a shared workspace
- The skill file is plain Markdown holding the run protocol, keyword selection rules and writing guidelines — bad drafts are usually fixed with one line
- Site-specific config lives in a separate prompt on the scheduled run: repo, audience, topic lanes, voice, delivery; headless scheduling, human merge gate
Per run: pull keyword candidates from DataForSEO (volume, difficulty, CPC) → read current ranking pages → draft → brief the media agent → open a draft PR on a seo/<slug> branch with post, images, index updates → email notification; a human merges or nothing ships.
Key lessons
- Their blog posts are TSX components that must build cleanly; once the prompt spelled out the repo layout, the three files to touch, style bans and the bun run build requirement, drafts arrived mergeable
- The writer doesn't need the biggest model when the method lives in the skill file
Honest bits: no posts have had time to rank yet, so ROI is still open; they're adding auto-written keyword rationale to PR bodies so rejecting takes 10 seconds.
Related event: Team Runs Two-Agent SEO Content Pipeline, 8 Draft PRs in Two Weeks(2 posts)→
More from coding & agent
- Simple trick to de-vibecode your app: feed the model a blacklist of AI-app cliches — mfckr_eth · 2026-09-02
- 'My 2T Token Life': Devs Show Off Agent Coding Token Guzzling With Tokscale — doodlestein · 2026-09-02
- Google Ships Gemini 3.8 Flash and Cyber Model for Agents and Code Security — GoogleDeepMind · 2026-09-02
- One Claude Code Skill Turns a 20-Shot Script Into 8 Finished AI Documentaries (33 Min) — illscience · 2026-09-02
- LangChain doubles down on model-agnostic stance: 'the right model today may not be right next week' — LangChain · 2026-09-02
- A useful categorization of the most common agent skill patterns — rseroter · 2026-09-02