Thinking AI: LinkedIn Post Generator

An advanced content generation tool that leverages an agentic AI workflow to create high-quality, engaging LinkedIn posts while streaming its reasoning in real-time.

PLay the Video

Overview

Creating consistently engaging content for professional platforms like LinkedIn is challenging and time-consuming. This tool goes beyond simple text generation by employing an agentic AI model that thinks, plans, and researches before writing.

You provide a topic, and the AI agent analyzes it, brainstorms unique angles, drafts two complete posts, and then enhances them with A/B testable hooks, data-driven hashtags, and authoritative citations. Best of all, you can watch the entire thought process unfold in the "Show thinking" panel, providing full transparency into the AI's strategy.

Core Features

This application is packed with features designed to streamline your content creation process. The table below outlines the key functionalities:

Feature	Description
Dynamic Post Generation	Simply enter a topic, and the AI generates two complete, distinct posts ready for publishing.
Customizable Voice & Style	Define your target Audience and Tone using presets or custom inputs. Provide a URL to a LinkedIn post, and the AI will mimic its writing style, including emoji usage, formatting, and sentence structure.
A/B Testable Hooks	Each post comes with multiple hook variations (questions and statements) organized in an accordion, allowing you to choose the perfect opener to maximize engagement.
Data-Driven Hashtags	The agent uses Google Search as a tool to find relevant and trending hashtags for your topic, increasing post visibility.
Authoritative Citations	To enhance credibility, the AI finds and links to recent, authoritative sources related to the post content.
Transparent AI Reasoning	A real-time stream shows you the AI's step-by-step plan, from scraping URLs to brainstorming angles and searching for hashtags.
Quality & Safety Guardrail	A built-in check reviews generated content for professionalism and appropriateness before it's presented to you.
API Health Check	The UI includes a service status indicator to confirm that the backend AI service is online and operational.
Seamless Export	Easily copy the post body, hashtags, or the entire formatted post with a single click, or save it as a .txt file.

The Agentic Workflow

What makes this tool special is its agentic AI workflow. Instead of executing a single, massive prompt, the AI acts as an autonomous agent that breaks down the complex task of "writing a good LinkedIn post" into a series of logical steps. It plans, uses tools, and refines its output to achieve a high-quality result.

Workflow Visualization

This diagram illustrates the step-by-step process the AI agent follows:

graph TD
    A[Start: User provides Topic, Tone, etc.] --> B{Style URL provided?};
    B -- Yes --> C[1. Scrape URL for Style Guide];
    B -- No --> D[1. Use Default Style];
    C --> E[1. Analyze & Create Style Guide];
    D --> E;
    E --> F[2. Brainstorm Post Angles];
    F --> G[3. Select Top 2 Angles];
    G --> H1[4a. Draft Post for Angle 1];
    G --> H2[4b. Draft Post for Angle 2];
    H1 --> I1[5a. Quality Guardrail Check];
    H2 --> I2[5b. Quality Guardrail Check];
    I1 --> J1[6a. Generate A/B Test Hooks];
    I2 --> J2[6b. Generate A/B Test Hooks];
    J1 --> K1[Google Search for Citations];
    J2 --> K2[Google Search for Citations];
    K1 --> L1[End: Assemble Post 1];
    K2 --> L2[End: Assemble Post 2];

Loading

Step-by-Step Breakdown

1. Understand Style: The agent first determines the desired writing style. If a URL is provided, it uses a web scraper (Cheerio) to fetch the content and then analyzes it to create a dynamic style guide. Otherwise, it uses the default tone specified by the user.
2. Plan the Attack (Angle Brainstorming): Before writing, the agent acts as a content strategist. It brainstorms several potential angles for the topic, considering the target audience, and then selects the two most unique and compelling ones to proceed with.
3. Draft Content: For each chosen angle, the agent drafts the main body of the post, focusing on creating valuable and engaging content that adheres to the style guide.
4. Quality & Safety Control: Each draft is passed through a guardrail check. The agent reviews the text for any unprofessional or inappropriate content. If any issues are found, it sanitizes the text; otherwise, it marks it as "APPROVED".
5. Refine & Enhance (Hook Generation): The agent generates multiple A/B testable hooks (both questions and statements) based on the opening line of each post, providing options to capture the audience's attention.
6. Research & Verify (Tool Use): In the final step, the agent uses Google Search as a tool to find:
   • Relevant and trending LinkedIn hashtags.
   • 2-3 recent, authoritative articles to use as citations, adding credibility and depth to the post.
7. Assemble & Deliver: The agent assembles all the generated components—angle, hooks, content, hashtags, and citations—into the final PostCard components displayed in the user interface.

Tech Stack

• Framework: Next.js (with Turbopack for faster builds)
• AI Model: Google Gemini Pro & Flash (for advanced reasoning and generation)
• Styling: Tailwind CSS (for responsive and modern UI)
• UI Components: shadcn/ui (built on Radix UI + Tailwind CSS for accessible components)
• Web Scraping: Cheerio (for extracting style from LinkedIn URLs)
• Deployment: Vercel (for seamless hosting and scaling)

Getting Started

Follow these steps to get the project running on your local machine.

Prerequisites

• Node.js (v18.18.0 or later)
• npm, yarn, or pnpm

1. Clone the Repository

git clone https://github.com/euclidstellar/linkedin-nugget.git
cd linkedin-nugget

2. Install Dependencies

npm install

3. Set Up Environment Variables

You'll need a Google Gemini API key to run this project.

1. Create a file named .env.local in the root of the project.

2. Get your API key from Google AI Studio.

3. Add the key to your .env.local file:

   GOOGLE_API_KEY="YOUR_API_KEY_HERE"

4. Run the Development Server

npm run dev

Open http://localhost:3000 in your browser to see the result.

Deploy on Vercel

The easiest way to deploy your Next.js app is to use the Vercel Platform from the creators of Next.js.

Don't forget to add your GOOGLE_API_KEY as an environment variable in your Vercel project settings.

Usage

1. Enter a topic in the input field.
2. Optionally, specify audience, tone, and provide a style URL.
3. Click "Generate Posts" to start the AI workflow.
4. Watch the real-time thinking stream in the "Show thinking" panel.
5. Review the generated posts, select hooks, and export as needed.

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

This project is licensed under the MIT License - see the LICENSE file for details.