LinkedIn Post Generator

​

Prerequisites: Setting Up LinkedIn MCP Server

​

Step 1: Configure LinkedIn MCP Server

1. Navigate to MCP Settings: Go to your Dume AI menu and locate the MCP Server section
2. Add LinkedIn Integration: Follow the detailed setup instructions at https://docs.dume.ai/mcp_server to connect LinkdIn MCP remote server from Composio.
3. Authenticate Your Account: Complete the OAuth flow to connect your LinkedIn profile
4. Test Connection: Verify that the LinkedIn MCP server is properly connected and accessible

​

Building the LinkedIn Post Workflow

​

Step 2: Create the Input Node

1. Add Input Node: Drag an Input Node onto your workflow canvas
2. Configure Input Parameters:
   • Field Name: post_topic (Type: Text, Required: Yes)
   • Field Name: tone (Type: Dropdown, Options: Professional, Casual, Enthusiastic, Required: Yes)
   • Field Name: target_audience (Type: Text, Required: No)
   • Field Name: key_points (Type: Text, Required: No)
3. Add Description: “Provide the topic and tone for your LinkedIn post”

​

Step 3: Create the Hook Generator LLM Node

1. Add LLM Node: Place an LLM Node and connect it to the Input Node
2. Node Configuration:
   • Name: “Hook Generator”
   • Model: Select “Dume AI Chat”
   • Prompt:

Create 2-3 engaging hook lines for a LinkedIn post about: {string INPUT/post_topic}

Tone: {string INPUT/tone}
Target Audience: {string INPUT/target_audience}

Make the hooks attention-grabbing and professional. Focus on creating curiosity or providing value.

3. Output Schema: Keep default (single string output)

​

Step 4: Create the Body Content Generator LLM Node

1. Add Second LLM Node: Place another LLM Node connected to the Input Node
2. Node Configuration:
   • Name: “Body Content Generator”
   • Model: Select “Dume AI Chat”
   • Prompt:

Write engaging LinkedIn post body content about: {string INPUT/post_topic}

Tone: {string INPUT/tone}
Key Points to Include: {string INPUT/key_points}
Target Audience: {string INPUT/target_audience}

Make it informative, engaging, and suitable for LinkedIn's professional audience. Keep it concise but valuable.

3. Output Schema: Default string output

​

Step 5: Create the Hashtag & CTA Generator LLM Node

1. Add Third LLM Node: Place another LLM Node connected to the Input Node
2. Node Configuration:
   • Name: “Hashtag & CTA Generator”
   • Model: Select “Dume AI Chat”
   • Prompt:

Generate 5-8 relevant hashtags and a compelling call-to-action for a LinkedIn post about: {string INPUT/post_topic}

Tone: {string INPUT/tone}
Target Audience: {string INPUT/target_audience}

Format:
Hashtags: #hashtag1 #hashtag2 #hashtag3...
CTA: [Your call-to-action here]

3. Output Schema: Default string output

​

Step 6: Create the Post Refiner LLM Node

1. Add Fourth LLM Node: Place an LLM Node that connects to all three previous LLM nodes
2. Node Configuration:
   • Name: “Post Refiner”
   • Model: Select “Dume AI Chat”
   • Prompt:

Combine and refine these LinkedIn post components into a cohesive, professional post:

Hook: {string Hook_Generator/answer}
Body: {string Body_Content_Generator/answer}  
Hashtags & CTA: {string Hashtag_CTA_Generator/answer}

Requirements:
- Ensure smooth flow between sections
- Optimize for LinkedIn's algorithm
- Keep professional tone
- Make sure the post is engaging and valuable
- Final post should be ready to publish

3. Output Schema: Default string output

​

Step 7: Add the LinkedIn Publishing Tool Node

1. Add Tool Node: Place a Tool Node connected to the Post Refiner
2. Node Configuration:
   • Name: “LinkedIn Publisher”
   • Integration: Select your configured LinkedIn MCP Server
   • Tool: Choose “Create LinkedIn Post” (or equivalent tool name)
   • Message/Instructions:

Publish this LinkedIn post: {string Post_Refiner/answer}

3. Parameters: The tool should automatically map the post content from the Post Refiner output

​

Step 8: Add the Output Node

1. Add Output Node: Place an Output Node connected to the Tool Node
2. Configure Output Variables:
   • Field Name: original_topic → Map to {string INPUT/post_topic}
   • Field Name: generated_post → Map to {string Post_Refiner/answer}
   • Field Name: publish_result → Map to {string LinkedIn_Publisher/result}
   • Field Name: post_url → Map to {string LinkedIn_Publisher/post_url} (if available)
3. Description: “LinkedIn post generation and publishing results”

​

Testing Your Workflow

​

Step 9: Test the Complete Flow

1. Save Your Workflow: Give it a descriptive name like “LinkedIn Post Creator”. It also auto saves in every 10 seconds.
2. Test Run: Click the test button and provide sample inputs:
   • Post Topic: “The future of AI in marketing”
   • Tone: “Professional”
   • Target Audience: “Marketing professionals and business owners”
   • Key Points: “Efficiency, personalization, data-driven insights”
3. Verify Results: Check that each node produces expected outputs and the final LinkedIn post is published successfully on the LinkedIn page.

​

Pro Tips

• Add Note Nodes: Include documentation for complex logic or team collaboration
• Error Handling: Consider adding Condition Nodes to handle potential API failures
• Customization: Adjust prompts based on your brand voice and industry
• Testing: Always test with different topics and tones before deploying