Description
This automated workflow generates social media posts using AI, sends them for human review via Slack, and publishes approved content to Twitter (X). It leverages Retrieval‑Augmented Generation (RAG) with context sourced from memory stores and vector search, ensuring high‑quality, context‑aware content with a built‑in approval step.

Overview

1. Trigger: Xpost Prompt starts the process with a content request.
2. RAG Agent: Enhances the prompt using relevant context from Postgres and a vector store.
3. Vector Search: Retrieves context from Pinecone using HuggingFace embeddings.
4. AI Generation: Generates a draft post with Google Gemini based on RAG results.
5. Slack Message: Sends the draft to Slack for human review.
6. AI Agent: Refines the draft further using memory and optional tools.
7. Post Validation: Sends the refined draft to Slack asking for approval.
8. Conditional Branch:
   • If approved: publishes the post to X (Twitter).
   • If rejected: sends feedback to Slack.

Tools Used

• n8n: Orchestration platform
• Google Gemini Chat Model: AI text generation
• Postgres Chat Memory: History and context storage
• HuggingFace Embeddings: Text-to-vector conversion
• Pinecone Vector Store: Semantic context retrieval
• Slack: Human-in-the-loop review and validation
• Twitter (X): Final post destination
• Custom AI Agents: Encapsulate RAG and workflow logic

Use Cases

• Brand Social Automation: Scale content creation with oversight
• Personal Thought Leadership: AI-generated insights with brand voice
• Content QA Workflows: Ensure quality before publishing
• Marketing Team Collaboration: Campaign-driven post creation and checks
• Startup Growth Loops: Automate social proof with intelligent content

Let me know if you’d like a version formatted for SEO, a headline and meta description, or any other adjustments!