AI Video Automation Workflow

n8n workflow that turns a content idea into a narrated video and a marketing campaign, built with LangChain, LangGraph and GoHighLevel.

RoleAI Automation Engineer

Timeline2026

TeamIndividual

Tech Stack8 Technologies

Mission Brief

An end-to-end AI video automation pipeline built around n8n, LangChain, LangGraph and GoHighLevel. A scheduled LangChain agent generates content ideas, a second agent turns them into detailed video prompts, and the workflow then generates the video and narration through the Fal API, assembles the final cut, and hands it to GoHighLevel for CRM updates, social scheduling and email/SMS campaigns. The focus is practical AI engineering: structured outputs, API integrations, asynchronous jobs with polling, and workflow state management.

Key Features

AI Content Pipeline

Idea generator: A scheduled LangChain agent with memory, tools and a structured output parser creates content ideas and logs them to GoHighLevel and Google Sheets.
Video prompts: An AI agent turns each idea into a detailed video prompt and updates the CRM opportunity.
Video and audio: Prompts go to the Fal API for video generation and text-to-speech, with polling until each job is ready and files stored in Google Drive.
Final assembly: Video and narration are merged through the Fal API, stored, and logged with their metadata.

Marketing Automation

CRM flow: Each idea moves through GoHighLevel stages: idea, prompts, video ready, published, campaign.
Distribution: Social posts are scheduled, contacts are added to email/SMS campaigns, and the team is notified by webhook through Slack or email.

Engineering Chronicles

Video and speech generation are slow, asynchronous jobs that can take minutes to finish.

SolutionQueue each job through the Fal API, poll its status with wait steps, and only continue once the file is ready, so every stage starts from a finished asset.