n8n Workflow Analysis: HeyGen AI Avatar Video Automation

This n8n workflow, authored by Nate Herk, is an automation designed to scrape daily news, synthesize a short video script using AI, and generate an AI-avatar video using the HeyGen API. It notably features a polling loop to monitor the video rendering status until completion.

Main Automation Flow (News to Video with Polling)

1. Manual Trigger: The workflow initiates via a manual Test workflow trigger.
2. Data Extraction (Apify Crawler): An HTTP Request node labeled News calls an Apify actor (website-content-crawler) to scrape the latest issue of the "Morning Brew" newsletter.
3. AI Script Generation (Langchain Agent): The scraped HTML/Markdown is passed to an AI Agent powered by GPT-4.1-mini (via OpenRouter). A specific system prompt instructs the AI to act as a sharp scriptwriter, summarizing the top 2-4 stories into a punchy, 10–20 second spoken script.
4. Video Generation (HeyGen API): A POST request is sent to HeyGen's /v2/video/generate endpoint. It specifies an avatar ID, a voice ID, and the text input. (Note: In the provided JSON, the text input is currently hardcoded as "Hey AI Automation Society Family. You guys rock.", but the architectural intent is to pipe the AI script here).
5. Rendering Polling Loop:
   • Delay: A Wait node pauses execution for 30 seconds.
   • Status Check: A GET request (Get Video1) fetches the rendering status of the generated video_id.
   • Condition (If Node): The workflow checks if the API response status equals "completed".
   • Feedback Loop: If the video is not yet completed, it triggers another Wait node and loops back to the Status Check until the render finishes.

Utility Nodes (Standalone API Calls)

The canvas includes a few isolated HTTP Request nodes used for setup and gathering necessary API variables:

• Get Avatars: Queries the HeyGen API to retrieve a list of available avatar_ids.
• Get Voices: Queries the HeyGen API to retrieve a list of available voice_ids.
• Standard Generate Video / Get Video: A simplified, two-node sequence for generating a video and manually checking its status without the automated polling loop.

Required Integrations & Credentials

To successfully run this workflow, the following services must be authenticated within n8n:

• HeyGen API Token: Configured as Header Auth for fetching avatars, voices, and triggering video generation.
• Apify API Token: Required in the News node body to run the web scraper.
• OpenRouter API Key: Required for the Langchain model to access OpenAI's GPT-4.1-mini.