# 🎬 HeyGen AI Avatar News Video Generator
[](https://n8n.io/)
[](https://heygen.com/)
[](https://ai.google.dev/)
[](https://github.com)
> **Fully automated AI video generator that transforms RSS news feeds into professional avatar-presented videos. Fetches headlines, generates scripts with Gemini AI, and creates videos with HeyGen's realistic AI avatars.**
Simply click execute and watch as your workflow automatically creates a news video from the latest BBC headlines—complete with an AI avatar presenter and professionally generated script.
----
🎥 **Project Demo:**
[](https://github.com/Awaisali36/ai-avatar-video-generation-system/blob/main/Vidoes%20/preview_video_target.mp4)
---
## 📋 Table of Contents
- [Overview](#-overview)
- [Key Features](#-key-features)
- [How It Works](#-how-it-works)
- [System Architecture](#-system-architecture)
- [Tech Stack](#-tech-stack)
- [Prerequisites](#-prerequisites)
- [Installation](#-installation)
- [Configuration](#-configuration)
- [Workflow Breakdown](#-workflow-breakdown)
- [Customization](#-customization)
- [Troubleshooting](#-troubleshooting)
- [Best Practices](#-best-practices)
- [License](#-license)
---
## 🎯 Overview
**HeyGen AI Avatar News Video Generator** is an intelligent n8n workflow that fully automates the creation of news videos. It combines RSS feeds, Google Gemini AI for script generation, and HeyGen's cutting-edge AI avatars to produce professional-quality videos without any manual editing.
### What Makes This Special?
- 🤖 **Fully Automated** - No manual intervention from feed to final video
- 📰 **RSS Integration** - Pulls latest news from any RSS feed (BBC News by default)
- 🧠 **AI Script Generation** - Gemini 2.5 Flash creates engaging news scripts
- 👤 **Realistic AI Avatars** - HeyGen's premium avatars present your content
- 🎙️ **Natural Voice Synthesis** - Automatic male voice selection
- ⏱️ **Smart Polling** - Waits for video generation with exponential backoff
- 📥 **Auto Download** - Saves completed video automatically
- 🔄 **Error Handling** - Robust retry logic and status checking
### Who Is This For?
- 📺 **Content Creators** - Generate news recap videos daily
- 🎥 **Social Media Managers** - Create engaging video content at scale
- 📱 **News Channels** - Automate news video production
- 🏢 **Marketing Teams** - Quick video content from industry news
- 🎓 **Educators** - Create educational news summaries
- 🤖 **Automation Enthusiasts** - Learn advanced n8n workflows
---
## ✨ Key Features
### 📰 **RSS Feed Integration**
- Fetches latest headlines from BBC News RSS
- Processes multiple articles automatically
- Creates digestible news summaries
- Supports any RSS feed URL
- Configurable number of headlines (default: 5)
### 🧠 **AI Script Generation (Google Gemini)**
- Uses Gemini 2.5 Flash model
- Converts headlines into engaging scripts
- Natural conversational tone
- Optimized for avatar delivery
- Character limit enforcement (1400 chars)
- Automatic script validation
### 🎬 **HeyGen Avatar Video Creation**
- Access to HeyGen's avatar library
- Automatic non-premium avatar selection
- Smart male voice picker
- Customizable video dimensions (1280x720)
- Professional avatar presentation
- Natural gestures and expressions
### 🎙️ **Voice Synthesis**
- Automatic male voice selection from HeyGen library
- Fallback voice ID for reliability
- Natural speech patterns
- Adjustable speech speed
- Gender preference matching
### ⏱️ **Smart Video Processing**
- Polls HeyGen API for completion status
- Exponential backoff strategy
- Maximum wait time protection
- Status checking (completed/failed/waiting/pending/processing)
- Automatic retry on transient failures
### 📥 **Video Download & Management**
- Automatic video file download
- Thumbnail and GIF URLs extracted
- Complete metadata capture
- Job ID tracking
- Duration and creation timestamps
---
## 🔄 How It Works
### User Experience
```
Step 1: Click "Execute Workflow" in n8n
Step 2: Workflow automatically:
├─ Fetches BBC News RSS feed
├─ Extracts top 5 headlines
├─ Creates news digest
├─ Sends to Gemini AI for script generation
├─ Fetches HeyGen avatars
├─ Fetches HeyGen voices
├─ Selects best avatar and voice
├─ Submits video generation request
├─ Polls for completion (every 6-30 seconds)
└─ Downloads finished video
Step 3: Video ready! (~2-5 minutes total)
• Professional AI avatar presentation
• Natural voice narration
• HD quality (1280x720)
• Ready to publish
```
### Technical Flow
```
Manual Trigger
↓
RSS Feed Reader (BBC News)
↓
Extract Top 5 Headlines
↓
Create News Digest
↓
Google Gemini API
├─ Generate engaging script
└─ Validate & truncate to 1400 chars
↓
Prepare Configuration
↓
Parallel Fetch:
├─ HeyGen Avatars → Pick Non-Premium Avatar
└─ HeyGen Voices → Pick Male Voice
↓
Generate Video Request (HeyGen API)
↓
Extract Video ID
↓
Poll Video Status Loop:
├─ Check Status
├─ Switch (completed/failed/waiting/pending/processing)
├─ If Still Processing:
│ ├─ Calculate Backoff Wait Time
│ ├─ Wait (6-30 seconds)
│ └─ Poll Again
└─ If Completed:
├─ Download Video File
├─ Extract Metadata
└─ Return Results
```
---
## 🏗️ System Architecture
### Data Flow
```
┌─────────────────────────────────────────────────────────────┐
│ INPUT LAYER │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ RSS Feed Reader │ │
│ │ (BBC News RSS) │ │
│ │ │ │
│ │ Fetches: Title, Description, Link, PubDate │ │
│ │ Returns: Array of news articles │ │
│ └────────────────────┬─────────────────────────┘ │
│ │ │
└───────────────────────┼──────────────────────────────────────┘
↓
┌─────────────────────────────────────────────────────────────┐
│ PROCESSING LAYER │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ Create News Digest │ │
│ │ (JavaScript Code Node) │ │
│ │ │ │
│ │ • Select first 5 headlines │ │
│ │ • Format as bullet points │ │
│ │ • Create single digest string │ │
│ └────────────────────┬─────────────────────────┘ │
│ │ │
│ ↓ │
│ ┌──────────────────────────────────────────────┐ │
│ │ Google Gemini API │ │
│ │ (Script Generation) │ │
│ │ │ │
│ │ Input: News digest │ │
│ │ Model: Gemini 2.5 Flash │ │
│ │ Output: Engaging news script │ │
│ └────────────────────┬─────────────────────────┘ │
│ │ │
│ ↓ │
│ ┌──────────────────────────────────────────────┐ │
│ │ Extract & Validate Script │ │
│ │ (JavaScript Code Node) │ │
│ │ │ │
│ │ • Extract text from Gemini response │ │
│ │ • Trim whitespace │ │
│ │ • Enforce 1400 character limit │ │
│ │ • Validate script is not empty │ │
│ └────────────────────┬─────────────────────────┘ │
│ │ │
└───────────────────────┼──────────────────────────────────────┘
↓
┌─────────────────────────────────────────────────────────────┐
│ HEYGEN PREPARATION │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ Set Default Configuration │ │
│ │ (Edit Fields Node) │ │
│ │ │ │
│ │ • Avatar ID: 9bb55566ca0242609aa1fea1f7144656│ │
│ │ • Voice ID: 0912d8004fd948a9aab02fffd264a3c1 │ │
│ │ • Dimensions: 1280x720 │ │
│ │ • Script: Pass through │ │
│ └────────────────────┬─────────────────────────┘ │
│ │ │
│ ┌────────────┴────────────┐ │
│ ↓ ↓ │
│ ┌──────────────────┐ ┌──────────────────┐ │
│ │ Fetch Avatars │ │ Fetch Voices │ │
│ │ (HeyGen API) │ │ (HeyGen API) │ │
│ └────────┬─────────┘ └────────┬─────────┘ │
│ ↓ ↓ │
│ ┌──────────────────┐ ┌──────────────────┐ │
│ │ Pick Avatar │ │ Pick Male Voice │ │
│ │ (Non-Premium) │ │ (Preference) │ │
│ └────────┬─────────┘ └────────┬─────────┘ │
│ │ │ │
│ └────────────┬────────────┘ │
│ │ │
└────────────────────────┼──────────────────────────────────────┘
↓
┌─────────────────────────────────────────────────────────────┐
│ VIDEO GENERATION │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ HeyGen Video Generation Request │ │
│ │ │ │
│ │ POST /v2/video/generate │ │
│ │ { │ │
│ │ "video_inputs": [{ │ │
│ │ "character": { avatar_id, avatar_style },│ │
│ │ "voice": { input_text, voice_id, speed } │ │
│ │ }], │ │
│ │ "dimension": { width, height } │ │
│ │ } │ │
│ └────────────────────┬─────────────────────────┘ │
│ │ │
│ ↓ │
│ ┌──────────────────────────────────────────────┐ │
│ │ Extract Video ID │ │
│ │ Initialize Polling State │ │
│ │ │ │
│ │ • video_id: From HeyGen response │ │
│ │ • attempts: 0 │ │
│ │ • max_poll_seconds: 180 │ │
│ │ • poll_interval_seconds: 6 │ │
│ └────────────────────┬─────────────────────────┘ │
│ │ │
└───────────────────────┼──────────────────────────────────────┘
↓
┌─────────────────────────────────────────────────────────────┐
│ POLLING LOOP │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ ┌─────────────────────────────────────┐ │ │
│ │ │ Poll Video Status │ │ │
│ │ │ GET /v1/video_status.get │ │ │
│ │ └────────────┬────────────────────────┘ │ │
│ │ │ │ │
│ │ ↓ │ │
│ │ ┌─────────────────────────────────────┐ │ │
│ │ │ Switch on Status │ │ │
│ │ ├─────────────────────────────────────┤ │ │
│ │ │ • completed → Download Video │ │ │
│ │ │ • failed → Throw Error │ │ │
│ │ │ • waiting → Continue Loop │ │ │
│ │ │ • pending → Continue Loop │ │ │
│ │ │ • processing → Continue Loop │ │ │
│ │ └────────────┬────────────────────────┘ │ │
│ │ │ │ │
│ │ ↓ (if not completed) │ │
│ │ ┌─────────────────────────────────────┐ │ │
│ │ │ Calculate Backoff Wait │ │ │
│ │ │ • Base: 6 seconds │ │ │
│ │ │ • Add: +2s per attempt │ │ │
│ │ │ • Max: 30 seconds │ │ │
│ │ │ • Check: Total time < 180s │ │ │
│ │ └────────────┬────────────────────────┘ │ │
│ │ │ │ │
│ │ ↓ │ │
│ │ ┌─────────────────────────────────────┐ │ │
│ │ │ Wait Node │ │ │
│ │ │ (Dynamic duration) │ │ │
│ │ └────────────┬────────────────────────┘ │ │
│ │ │ │ │
│ │ └──────────┐ │ │
│ │ │ │ │
│ └──────────────────────────┼────────────────────┘ │
│ │ (Loop back) │
└─────────────────────────────┼──────────────────────────────────┘
↓ (completed)
┌─────────────────────────────────────────────────────────────┐
│ OUTPUT LAYER │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ Download Video File │ │
│ │ (HTTP Request - Binary Response) │ │
│ └────────────────────┬─────────────────────────┘ │
│ │ │
│ ↓ │
│ ┌──────────────────────────────────────────────┐ │
│ │ Format Output Metadata │ │
│ │ │ │
│ │ • video_url: Direct download link │ │
│ │ • thumbnail_url: Preview image │ │
│ │ • gif_url: Animated preview │ │
│ │ • job_id: HeyGen job identifier │ │
│ │ • status: completed │ │
│ │ • created_at: Timestamp │ │
│ │ • duration: Video length │ │
│ │ • message: Success confirmation │ │
│ └──────────────────────────────────────────────┘ │
│ │
└─────────────────────────────────────────────────────────────┘
```
---
## 🛠️ Tech Stack
| Category | Technology | Purpose |
|----------|------------|---------|
| **Automation** | n8n | Workflow orchestration |
| **RSS Feed** | BBC News RSS | News headline source |
| **AI Script** | Google Gemini 2.5 Flash | Script generation |
| **Avatar Platform** | HeyGen | AI avatar video creation |
| **Voice Synthesis** | HeyGen Voices | Natural speech generation |
| **Programming** | JavaScript (n8n Code nodes) | Data processing and logic |
| **API Integration** | HTTP Request nodes | External service communication |
---
## 📦 Prerequisites
### Required Accounts & API Keys
| Service | Required? | Purpose | Cost |
|---------|-----------|---------|------|
| **n8n** | ✅ Yes | Run workflows | Free (self-hosted) or $20/mo |
| **Google AI Studio** | ✅ Yes | Gemini API access | Free tier: 15 requests/min |
| **HeyGen** | ✅ Yes | AI avatar videos | $24-$120/mo (based on plan) |
### API Keys Needed
- ✅ Google Gemini API Key
- ✅ HeyGen API Key (X-Api-Key)
### Technical Requirements
- n8n instance (v1.0+)
- Stable internet connection
- ~5 GB storage for video files
---
## 🚀 Installation
### Step 1: Download Workflow
1. Download `My workflow.json` from repository
2. Save to your computer
### Step 2: Import to n8n
1. Open your n8n instance
2. Click **"Workflows"** → **"Import from File"**
3. Select `My workflow.json`
4. Click **"Import"**
5. Workflow appears in your workspace
### Step 3: Get Google Gemini API Key
1. Visit https://ai.google.dev/
2. Click **"Get API key"**
3. Create new project or select existing
4. Click **"Create API Key"**
5. Copy and save the key
**Format:** `AIzaSyAxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx`
### Step 4: Get HeyGen API Key
1. Sign up at https://www.heygen.com/
2. Go to **"API"** section in dashboard
3. Copy your API key
**Format:** `MDU2NTExODlmZWJmNGY2MTk4ZjBhMjEzMjM1NGYyNGItMTc1OTY0NjA2Ng==`
**Note:** The key in the workflow is the actual HeyGen key from the JSON. Replace it with your own key!
---
## ⚙️ Configuration
### Update API Keys
You need to update API keys in **5 HTTP Request nodes**:
#### 1. HTTP Request20 (Gemini API)
Find this node and update the URL:
**Before:**
```
https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash:generateContent?key=AIzaSyAqNnJtQE888wIZHoTCYwP1pn0sZ3X2y_Q
```
**After:**
```
https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash:generateContent?key=YOUR_GEMINI_API_KEY
```
---
#### 2. HTTP Request2 (Fetch Avatars)
Find the node and update headers:
**Header Parameters:**
- Name: `X-Api-Key`
- Value: Replace with `YOUR_HEYGEN_API_KEY`
---
#### 3. HTTP Request16 (Fetch Voices)
**Header Parameters:**
- Name: `X-Api-Key`
- Value: Replace with `YOUR_HEYGEN_API_KEY`
---
#### 4. HTTP Request17 (Generate Video)
**Header Parameters:**
- Name: `x-api-key`
- Value: Replace with `YOUR_HEYGEN_API_KEY`
---
#### 5. HTTP Request18 (Poll Status)
**Header Parameters:**
- Name: `X-Api-Key`
- Value: Replace with `YOUR_HEYGEN_API_KEY`
---
### Customize RSS Feed
By default, workflow uses BBC News. To change:
1. Find **"RSS Read2"** node
2. Update URL parameter:
**Current:**
```
https://feeds.bbci.co.uk/news/rss.xml
```
**Other options:**
```
CNN: http://rss.cnn.com/rss/cnn_topstories.rss
Reuters: http://feeds.reuters.com/reuters/topNews
TechCrunch: https://techcrunch.com/feed/
```
---
### Adjust Headline Count
1. Find **"Code1"** node (Create Digest)
2. Modify `maxItems` value:
**Current:**
```javascript
const maxItems = 5;
```
**Change to:**
```javascript
const maxItems = 10; // Get 10 headlines instead
```
---
### Customize Video Settings
1. Find **"Edit Fields1"** node
2. Adjust values:
**Avatar ID:**
```javascript
"avatar_id": "9bb55566ca0242609aa1fea1f7144656"
```
**Voice ID:**
```javascript
"voice_id": "0912d8004fd948a9aab02fffd264a3c1"
```
**Video Dimensions:**
```javascript
"width": 1920, // Change to 1920 for Full HD
"height": 1080 // Change to 1080 for Full HD
```
**Default Script** (for testing):
```javascript
"script": "Your custom test script here"
```
---
### Polling Configuration
The workflow uses smart polling with exponential backoff. To adjust:
1. Find **"Code15"** node (Initialize polling)
2. Modify default values:
**Current:**
```javascript
max_poll_seconds: 180, // 3 minutes max
poll_interval_seconds: 6 // Start at 6 seconds
```
**For faster polling:**
```javascript
max_poll_seconds: 300, // 5 minutes max
poll_interval_seconds: 4 // Start at 4 seconds
```
---
## 📖 Workflow Breakdown
### Stage 1: News Collection
**Nodes: Manual Trigger → RSS Read2 → Code1**
1. **Manual Trigger**: Click to start workflow
2. **RSS Read2**: Fetches BBC News RSS feed
3. **Code1 (Create Digest)**:
```javascript
// Selects first 5 headlines
const maxItems = 5;
const selected = input.slice(0, maxItems);
// Formats as bullet points
const lines = selected.map(it => `• ${it.json.title}`);
const digest = lines.join('\n');
```
**Output Example:**
```
• UK inflation rises to 2.6% amid energy costs
• Climate summit reaches historic agreement
• Tech giant announces new AI breakthrough
• Sports: Manchester United wins championship
• Markets surge on positive economic data
```
---
### Stage 2: Script Generation
**Nodes: HTTP Request20 → Code14**
1. **HTTP Request20 (Gemini API)**:
- Sends news digest to Gemini 2.5 Flash
- Model generates engaging script
- Returns JSON response
2. **Code14 (Extract Script)**:
```javascript
// Extract text from Gemini response
const cands = items[0].json.candidates || [];
const text = cands?.[0]?.content?.parts?.[0]?.text || "";
// Safety: trim & hard cap at 1400 chars
const script = text.trim().slice(0, 1400);
if (!script) {
throw new Error("Empty script from Gemini");
}
```
**Output Example:**
```
"Good morning! Let's dive into today's top stories.
Starting with the economy, UK inflation has climbed to 2.6%..."
```
---
### Stage 3: HeyGen Preparation
**Nodes: Use My Script1 → Edit Fields1 → HTTP Request2 & HTTP Request16**
1. **Use My Script1**: Assigns script to variable
2. **Edit Fields1**: Sets default configuration
3. **Parallel Execution**:
- **HTTP Request2**: Fetches available avatars
- **HTTP Request16**: Fetches available voices
---
### Stage 4: Avatar & Voice Selection
**Nodes: Pick Avatar1 → Pick Male Voice1**
**Pick Avatar1 (JavaScript)**:
```javascript
// Choose first non-premium avatar
const avatars = root?.data?.avatars || root?.avatars || [];
const chosen = avatars.find(a => a?.premium === false) || avatars[0];
const avatarId = chosen.avatar_id || 'Abigail_expressive_2024112501';
```
**Pick Male Voice1 (JavaScript)**:
```javascript
// Prefer explicit male gender
const voices = root?.data?.voices || root?.voices || [];
const preferred = voices.find(v =>
String(v.gender || '').toLowerCase() === 'male'
);
// Fallback if no male voice found
const voiceId = preferred?.voice_id ?? '06f9223be7d444d887045ba89c1d00b6';
```
---
### Stage 5: Video Generation
**Node: HTTP Request17**
**API Request:**
```javascript
POST https://api.heygen.com/v2/video/generate
Headers:
- x-api-key: YOUR_HEYGEN_API_KEY
- Content-Type: application/json
Body:
{
"video_inputs": [{
"character": {
"type": "avatar",
"avatar_id": "{{ selected_avatar_id }}",
"avatar_style": "normal"
},
"voice": {
"type": "text",
"input_text": "{{ script }}",
"voice_id": "{{ selected_voice_id }}",
"speed": 1.1
}
}],
"dimension": {
"width": 1280,
"height": 720
}
}
```
**Response:**
```json
{
"data": {
"video_id": "abc123xyz789"
}
}
```
---
### Stage 6: Status Polling Loop
**Nodes: Code15 → HTTP Request18 → Switch2 → Code16 → Wait2**
**Flow:**
1. **Code15**: Extract video_id, initialize polling state
2. **HTTP Request18**: Check video status
3. **Switch2**: Route based on status
- **completed** → Download video
- **failed** → Throw error
- **waiting/pending/processing** → Continue loop
4. **Code16**: Calculate backoff wait time
```javascript
const prev = $json.attempts || 0;
const base = $json.poll_interval_seconds || 6;
// Exponential backoff: 6s, 8s, 10s, 12s, ..., max 30s
const wait = Math.min(base + prev * 2, 30);
// Check if exceeded max polling time
if (elapsed >= 180) {
throw new Error("Polling budget exceeded");
}
```
5. **Wait2**: Pause for calculated duration
6. **Loop back** to HTTP Request18
---
### Stage 7: Video Download
**Nodes: If2 → HTTP Request19 → Edit Fields5**
1. **If2**: Verify video_url exists
2. **HTTP Request19**: Download video file
- Response format: Binary (video file)
3. **Edit Fields5**: Format metadata
```javascript
{
"video_url": "https://...",
"thumbnail_url": "https://...",
"gif_url": "https://...",
"job_id": "...",
"status": "completed",
"duration": "...",
"message": "Video successfully generated"
}
```
---
## 🎨 Customization
### Use Different News Sources
**Add Multiple RSS Feeds:**
1. Duplicate **RSS Read2** node
2. Update URL for each source:
```
Source 1: BBC News
Source 2: CNN
Source 3: TechCrunch
```
3. Merge results before **Code1** node
---
### Customize Script Prompt
Add system prompt to Gemini for specific styles:
**In HTTP Request20 body:**
```javascript
{
"contents": [{
"parts": [{
"text": "You are a professional news anchor. Create an engaging 60-second news script from these headlines. Use a conversational yet authoritative tone.\n\nHeadlines:\n" + $json.digest
}]
}]
}
```
**Prompt variations:**
- "Create a humorous take on these news stories"
- "Write in the style of a sports commentator"
- "Make it suitable for children aged 8-12"
---
### Select Specific Avatar
Instead of auto-selection:
**In Edit Fields1 node:**
```javascript
"avatar_id": "YOUR_SPECIFIC_AVATAR_ID"
```
**To find avatar IDs:**
1. Run workflow once
2. Check output of HTTP Request2
3. Browse `data.avatars` array
4. Copy desired `avatar_id`
---
### Choose Female Voice
**In Pick Male Voice1 node**, change logic:
```javascript
// Change from male to female
const preferred = voices.find(v =>
String(v.gender || '').toLowerCase() === 'female'
);
// Fallback female voice
const voiceId = preferred?.voice_id ?? 'FEMALE_VOICE_ID';
```
---
### Add Background Music
This requires HeyGen Pro features. In video generation request:
```javascript
{
"video_inputs": [...],
"dimension": {...},
"audio": {
"type": "music",
"url": "https://yourmusicurl.com/background.mp3",
"volume": 0.3
}
}
```
---
### Schedule Automatic Execution
Replace **Manual Trigger** with **Schedule Trigger**:
1. Delete "When clicking 'Execute workflow'" node
2. Add **"Schedule Trigger"** node
3. Configure schedule:
```
Cron Expression: 0 8 * * *
Meaning: Every day at 8:00 AM
```
**Other schedules:**
```
Every hour: 0 * * * *
Every 6 hours: 0 */6 * * *
Weekdays at 9 AM: 0 9 * * 1-5
```
---
## 🐛 Troubleshooting
### Issue: "Empty script from Gemini"
**Cause:** Gemini API returned no text
**Solutions:**
1. Check API key is valid
2. Verify API quota not exceeded
3. Check if digest was properly formatted
4. Test Gemini API directly
5. Review request body in HTTP Request20
---
### Issue: "HeyGen did not return video_id"
**Cause:** Video generation request failed
**Solutions:**
1. Verify HeyGen API key is correct
2. Check HeyGen account has credits
3. Verify avatar_id is valid
4. Confirm voice_id exists
5. Check script length (<1400 chars)
6. Review HTTP Request17 response
---
### Issue: "Polling budget exceeded"
**Cause:** Video took longer than 3 minutes
**Solutions:**
1. Increase `max_poll_seconds` in Code15:
```javascript
max_poll_seconds: 300 // 5 minutes
```
2. Check HeyGen service status
3. Try shorter script
4. Verify account limits
---
### Issue: Video quality is poor
**Cause:** Resolution settings or HeyGen plan limits
**Solutions:**
1. Upgrade HeyGen plan for better quality
2. Increase dimensions in Edit Fields1:
```javascript
"width": 1920,
"height": 1080
```
3. Select premium avatars (may cost more)
---
### Issue: Wrong avatar selected
**Cause:** Auto-selection picking incorrect avatar
**Solutions:**
1. Hardcode specific avatar ID
2. Modify Pick Avatar1 logic:
```javascript
// Prefer specific style
const chosen = avatars.find(a =>
a.avatar_style === 'professional'
);
```
3. Review available avatars list
---
### Issue: Voice sounds robotic
**Cause:** Voice selection or speed settings
**Solutions:**
1. Try different voice_id
2. Adjust speech speed in HTTP Request17:
```javascript
"speed": 1.0 // Slower, more natural
```
3. Select premium voices
4. Check voice preview on HeyGen
---
### Issue: RSS feed fails to load
**Cause:** Feed URL changed or network issue
**Solutions:**
1. Test RSS URL in browser
2. Check for redirects
3. Try alternative RSS feeds
4. Add error handling:
```javascript
if (!items || items.length === 0) {
throw new Error("No RSS items found");
}
```
---
## 💡 Best Practices
### Cost Management
**HeyGen Credits:**
- Each video consumes credits
- Monitor credit usage in dashboard
- Consider bulk credit purchases
- Use non-premium avatars for testing
**Gemini API:**
- Free tier: 15 requests/minute
- Cache script results if re-using
- Monitor quota usage
**Cost per video (estimated):**
```
Gemini API: $0.00 (free tier)
HeyGen: $0.50 - $2.00 per video
Total: ~$0.50 - $2.00 per video
```
---
### Script Optimization
**Best practices:**
- Keep scripts under 60 seconds (~150 words)
- Use conversational language
- Avoid complex terms unless necessary
- Include pauses (commas, periods)
- Test different prompts
**Example good script:**
```
"Good morning! In today's news... [pause]
The UK economy shows mixed signals with inflation at 2.6%.
Meanwhile, climate leaders reached a historic agreement..."
```
---
### Avatar Selection
**Guidelines:**
- Use professional avatars for news
- Match avatar ethnicity to audience
- Consider gender based on voice
- Test different avatars for engagement
**Popular avatar types:**
- News anchor style
- Business professional
- Casual presenter
---
### Performance Optimization
**Speed up workflow:**
1. Reduce headline count (5 → 3)
2. Shorter scripts generate faster
3. Use standard resolution (720p vs 1080p)
4. Consider parallel processing for multiple videos
---
### Error Recovery
**Add error handling:**
**After Gemini request:**
```javascript
// Retry logic
let attempts = 0;
while (attempts < 3) {
try {
// API call
break;
} catch (e) {
attempts++;
if (attempts === 3) throw e;
// Wait and retry
}
}
```
---
### Quality Assurance
**Before publishing videos:**
- Watch complete video
- Check audio sync
- Verify script accuracy
- Test on target platform
- Review thumbnail/preview
---
## 📊 Use Cases
### 1. Daily News Recaps
**Schedule:**
- Run every morning at 8 AM
- Generate 2-minute news summary
- Post to YouTube/social media
**Benefits:**
- Consistent content schedule
- No video editing required
- Professional presentation
---
### 2. Industry News Updates
**Example: Tech News**
- RSS: TechCrunch
- Script focus: Latest tech developments
- Avatar: Professional tech presenter
**Distribution:**
- LinkedIn posts
- Company newsletter
- Internal communications
---
### 3. Educational Content
**Example: Daily History Facts**
- RSS: History feed
- Script: Engaging historical narratives
- Avatar: Teacher/educator style
**Target:**
- Educational platforms
- Social media education accounts
---
### 4. Market Updates
**Example: Financial News**
- RSS: Bloomberg/Reuters
- Script: Market analysis summary
- Avatar: Business professional
**Audience:**
- Investors
- Business professionals
- Financial advisors
---
### 5. Social Media Content
**Quick video clips:**
- 30-second news highlights
- Shareable format
- Optimized for mobile
**Platforms:**
- Instagram Reels
- TikTok
- YouTube Shorts
- Twitter/X
---
## 📈 Performance Metrics
**Workflow Execution Time:**
| Stage | Duration |
|-------|----------|
| RSS Feed Fetch | 2-5 seconds |
| Script Generation | 3-8 seconds |
| Avatar/Voice Selection | 2-4 seconds |
| Video Generation Request | 1-2 seconds |
| Video Processing | 60-180 seconds |
| Video Download | 5-15 seconds |
| **Total** | **~2-5 minutes** |
**Resource Usage:**
- CPU: Low (mostly API calls)
- Memory: <100 MB
- Network: ~50-200 MB per video
- Storage: 10-50 MB per video
---
## 💰 Cost Analysis
### Monthly Costs (30 videos/month)
| Service | Cost |
|---------|------|
| **n8n** (self-hosted) | $0 or $20/mo (cloud) |
| **Google Gemini** | $0 (free tier) |
| **HeyGen** | $24-120/mo (plan-based) |
| **Total** | **$24-140/month** |
**Cost per video:** $0.80 - $4.67
---
## 🚀 Advanced Features
### Multi-Language Support
Generate videos in different languages:
**Modify Gemini prompt:**
```javascript
"Generate this news script in Spanish..."
```
**Select appropriate voice:**
- Filter voices by language in Pick Male Voice node
---
### A/B Testing
Create multiple versions:
1. Generate with different avatars
2. Try various script styles
3. Test speech speeds
4. Compare engagement metrics
---
### Batch Processing
Generate multiple videos:
1. Add loop node before RSS Read
2. Array of topics/sources
3. Process each sequentially
4. Save all videos
---
## 📄 License
This project is licensed under the **MIT License**.
✅ Commercial use allowed
✅ Modification allowed
✅ Distribution allowed
✅ Private use allowed
⚠️ No warranty or liability
---
## 🌟 Acknowledgments
Built with powerful tools:
- [n8n](https://n8n.io) - Workflow automation
- [HeyGen](https://heygen.com) - AI avatar platform
- [Google Gemini](https://ai.google.dev/) - Script generation AI
- [BBC News RSS](https://www.bbc.com/news) - News source
---
## 🎬 Ready to Create AI Videos!
**Start generating professional AI avatar videos automatically:**
1. ✅ Import workflow to n8n
2. ✅ Add API keys (Gemini & HeyGen)
3. ✅ Click execute
4. ✅ Wait 2-5 minutes
5. ✅ Download your video!
**Made with 🎬 for content creators everywhere**
⭐ Automate your video content creation!
🤖 Professional AI avatars at scale!
📺 High-quality videos without editing!
---
