This project provides a LangGraph-based agent for sourcing, curating, and scheduling social media posts, with a human-in-the-loop workflow for review and approval. It targets content creators and social media managers looking to automate post generation from URLs, offering a flexible setup from basic to advanced features.

How It Works

The agent leverages LangGraph to orchestrate a multi-step process. It takes a URL, uses FireCrawl for web scraping, and then employs an LLM (Anthropic) to generate draft posts for Twitter and LinkedIn. A key aspect is the human-in-the-loop (HITL) mechanism, facilitated by Arcade for authentication and an optional Agent Inbox for review, allowing users to approve, reject, or edit generated content before publishing.

Quick Start & Requirements

• Install: yarn install
• Run: yarn langgraph:in_mem:up then yarn generate_post
• Prerequisites: Anthropic API key, FireCrawl API key, Arcade API key. Optional: LangSmith API key, Twitter Developer Account, LinkedIn Developer Account, Supabase account, GitHub API token, Slack Developer Account.
• Setup: Requires cloning the repo, installing Node.js dependencies (yarn), and setting environment variables.
• Docs: Video Tutorial, Full Setup Guide

Highlighted Details

• Supports posting to LinkedIn as an individual or organization.
• Integrates with Slack for data ingestion and status updates.
• Enables media uploads to Twitter via Twitter Developer Account setup.
• Offers an Agent Inbox for managing and reviewing generated posts.

Maintenance & Community

• Developed by LangChain AI.
• Links to Agent Inbox for UI interaction.

Licensing & Compatibility

• License: Not explicitly stated in the README.
• Compatibility: Requires Node.js and various API keys.

Limitations & Caveats

The "basic setup mode" omits features like parsing content from GitHub, Twitter, or YouTube, ingesting data from Slack, and image selection/uploads. Advanced features require significant API key setup and developer account configurations for multiple platforms.