AI News Pipeline Automation For Wordpress News Site
Budget / Salary$30–250
TypeFreelance project
LocationRemote
Posted1 hour ago
I want to automate the entire flow of finding, rewriting, and posting breaking stories in my niche WordPress newsroom. The system needs to:
• Continuously watch both Google News and selected RSS feeds for fresh headlines that match a keyword list I will supply.
• Pull the full-text of each candidate story, strip boilerplate, and hand it to an LLM (GPT-4o, Claude, or similar) to rewrite with a clear, human editorial touch—no robotic phrasing, no plagiarism.
• Automatically add an introductory line that credits the original outlet before the body copy begins.
• Push the finished piece to WordPress through the REST API, setting the post to Draft or Publish based on a simple flag in the configuration.
Ideal workflow: lightweight Python (or Node) microservice running on a VPS, cron or webhook triggered, Redis/SQLite for de-duping, and clean modular code so I can extend it later (extra feeds, different post types, etc.).
Acceptance criteria:
1. End-to-end demo where a new story breaking on Google News or an RSS feed appears as a draft in WordPress within five minutes.
2. Rewritten copy passes originality checks and reads naturally.
3. Intro sentence clearly names the original source every time.
4. Config file lets me toggle draft/publish, add feeds, and adjust keyword filters without touching code.
News Monitoring & Scraping: Set up reliable Google News RSS parameters and targeted source feeds to detect breaking news in real-time.
Duplicate Prevention: Implement logic (hash checks, URL tracking, or vector similarity) to ensure the system never processes or publishes duplicate stories.
High-Quality Human-Style AI Rewriting: Construct a master prompt/pipeline that rewrites the news using an Inverted Pyramid newsroom structure. The output must avoid AI clichés/buzzwords, sound like a professional beat reporter, and maintain strict factual accuracy.
Source Attribution & Credit: Enforce strict rules within the AI prompt to give full, professional credit to the original breaking source (e.g., "As first reported by [Outlet]...") both in the body text and in a formatted HTML source box at the end of the post.
WordPress Integration: Automatically output ready-to-publish HTML posts complete with formatted headlines, subheaders, categories, tags, and featured image handling via the WordPress REST API or native plugin integration.
Error Handling & Quality Gate: Configure an initial review setup (saving as Draft / Pending Review) with automated alerts (Slack/Email) if an API call or feed fails.
• Continuously watch both Google News and selected RSS feeds for fresh headlines that match a keyword list I will supply.
• Pull the full-text of each candidate story, strip boilerplate, and hand it to an LLM (GPT-4o, Claude, or similar) to rewrite with a clear, human editorial touch—no robotic phrasing, no plagiarism.
• Automatically add an introductory line that credits the original outlet before the body copy begins.
• Push the finished piece to WordPress through the REST API, setting the post to Draft or Publish based on a simple flag in the configuration.
Ideal workflow: lightweight Python (or Node) microservice running on a VPS, cron or webhook triggered, Redis/SQLite for de-duping, and clean modular code so I can extend it later (extra feeds, different post types, etc.).
Acceptance criteria:
1. End-to-end demo where a new story breaking on Google News or an RSS feed appears as a draft in WordPress within five minutes.
2. Rewritten copy passes originality checks and reads naturally.
3. Intro sentence clearly names the original source every time.
4. Config file lets me toggle draft/publish, add feeds, and adjust keyword filters without touching code.
News Monitoring & Scraping: Set up reliable Google News RSS parameters and targeted source feeds to detect breaking news in real-time.
Duplicate Prevention: Implement logic (hash checks, URL tracking, or vector similarity) to ensure the system never processes or publishes duplicate stories.
High-Quality Human-Style AI Rewriting: Construct a master prompt/pipeline that rewrites the news using an Inverted Pyramid newsroom structure. The output must avoid AI clichés/buzzwords, sound like a professional beat reporter, and maintain strict factual accuracy.
Source Attribution & Credit: Enforce strict rules within the AI prompt to give full, professional credit to the original breaking source (e.g., "As first reported by [Outlet]...") both in the body text and in a formatted HTML source box at the end of the post.
WordPress Integration: Automatically output ready-to-publish HTML posts complete with formatted headlines, subheaders, categories, tags, and featured image handling via the WordPress REST API or native plugin integration.
Error Handling & Quality Gate: Configure an initial review setup (saving as Draft / Pending Review) with automated alerts (Slack/Email) if an API call or feed fails.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.