graph TD %%{init: {'theme': 'mc','layout': 'elk'}}%% TextInput-st0k1[<div><img src="/_astro/type.Dy26vmDy.svg" style="height: 20px !important;width: 20px !important"/></div>Query] style TextInput-st0k1 stroke:#a170ff Prompt-2zkyi[<div><img src="/_astro/square-terminal.BMOXc-nZ.svg" style="height: 20px !important;width: 20px !important"/></div>Prompt] style Prompt-2zkyi stroke:#a170ff SearXng-q92z6[Web Search SearXng] style SearXng-q92z6 stroke:#a170ff WebScraper-wulwf[Web Scraper] style WebScraper-wulwf stroke:#a170ff Prompt-ctj5v[<div><img src="/_astro/square-terminal.BMOXc-nZ.svg" style="height: 20px !important;width: 20px !important"/></div>Prompt] style Prompt-ctj5v stroke:#a170ff ParseData-11ts5[<div><img src="/_astro/braces.Djq0PW4_.svg" style="height: 20px !important;width: 20px !important"/></div>Parse Data2] style ParseData-11ts5 stroke:#a170ff OpenAIModel-7fyyb[<div><img src="/_astro/openAI.BhmuxEs3.svg" style="height: 20px !important;width: 20px !important"/></div>OpenAI] style OpenAIModel-7fyyb stroke:#a170ff OpenAIModel-1p9j6[<div><img src="/_astro/openAI.BhmuxEs3.svg" style="height: 20px !important;width: 20px !important"/></div>OpenAI] style OpenAIModel-1p9j6 stroke:#a170ff ParseData-mo9ts[<div><img src="/_astro/braces.Djq0PW4_.svg" style="height: 20px !important;width: 20px !important"/></div>Parse Data] style ParseData-mo9ts stroke:#a170ff XPoster-usgch[X Poster] style XPoster-usgch stroke:#a170ff TextInput-st0k1 -.- SearXng-q92z6 linkStyle 0 stroke:#a170ff SearXng-q92z6 -.- ParseData-11ts5 linkStyle 1 stroke:#a170ff ParseData-11ts5 -.- Prompt-ctj5v linkStyle 2 stroke:#a170ff Prompt-ctj5v -.- OpenAIModel-7fyyb linkStyle 3 stroke:#a170ff OpenAIModel-7fyyb -.- WebScraper-wulwf linkStyle 4 stroke:#a170ff Prompt-2zkyi -.- OpenAIModel-1p9j6 linkStyle 5 stroke:#a170ff WebScraper-wulwf -.- ParseData-mo9ts linkStyle 6 stroke:#a170ff OpenAIModel-7fyyb -.- Prompt-2zkyi linkStyle 7 stroke:#a170ff ParseData-mo9ts -.- Prompt-2zkyi linkStyle 8 stroke:#a170ff OpenAIModel-1p9j6 -.- XPoster-usgch linkStyle 9 stroke:#a170ff

Article Tweet Generator

🧩 Overview

The workflow transforms a user’s query about recent AI news into a concise, tweet‑ready statement. It automatically searches the web, selects a relevant article, extracts its content, generates a tweet that summarizes the key points, and posts it on X. The process is fully automated, reducing manual curation and publishing time.

⚙️ Main Features

Accepts natural‑language queries about AI news.
Retrieves recent articles through a web search.
Randomly selects one article URL using a language model.
Scrapes the full article content.
Generates an informative, character‑limited tweet.
Publishes the tweet on X with a single automated step.

🔄 Workflow Steps

Component Name	Role in the Workflow	Key Inputs	Key Outputs
Query	Receives the user’s request for AI news.	User‑provided text.	Search query string.
Web Search	Conducts a web search using the query.	Search query string.	Data containing search result URLs.
Parse Data (Extract URLs)	Converts search results into a list of article URLs.	Search result data.	List of URLs.
Prompt (Select Random Article)	Asks the model to choose a single article URL.	List of article URLs.	Prompt string for the language model.
OpenAI Model (Generate URL)	Executes the model to output the selected URL.	Prompt string.	Selected article URL.
Web Scraper	Scrapes the article at the given URL.	Article URL.	Raw article content in Markdown.
Parse Data (Process Article)	Transforms scraped Markdown into plain text.	Raw article content.	Article content text.
Prompt (Generate Tweet)	Constructs a prompt for creating a tweet from the article.	Article content text, Article URL.	Prompt string for the language model.
OpenAI Model (Generate Tweet Text)	Generates the final tweet text.	Prompt string.	Tweet text (≤ 100 characters).
X Poster	Publishes the tweet to X.	Tweet text.	Confirmation of posted tweet.

🧠 Notes

API Credentials: The workflow requires valid API keys for both OpenAI and the X API.
Model Selection: Two language‑model calls are made—one for URL selection and one for tweet generation.
Randomness: The URL selection is performed by the LLM, which may produce different results on each run; consider adding a deterministic seed if reproducibility is needed.
Scraping Constraints: The web scraper’s timeout can be adjusted; large pages may exceed the default limit.
Tweet Length: The prompt explicitly restricts the output to 100 characters to comply with X’s character limit.
Safety: The system does not validate the content of the selected article before posting; additional filtering or moderation may be added if necessary.