Create Images, Edit Photos, Write Posts, and Generate Videos – All No Code

A companion guide to Nate Herk’s video tutorial (Free n8n templates included)

What if you could have an entire marketing team that creates images, edits photos, writes blog posts, generates LinkedIn content, and even produces videos – all controlled by a single AI agent you can talk to through Telegram?

That’s exactly what this tutorial builds. One AI agent with six powerful tools, all running on n8n with no coding required. And the best part? All the templates are available for free.

Video Chapters

Jump to any section:

What You’re Building

One AI agent that you communicate with through Telegram (voice or text) that has access to six powerful tools:

1. Create Image – Generate images using OpenAI’s new GPT-4o image model
2. Edit Image – Modify existing images while keeping them consistent
3. Search Image Database – Find and retrieve previously created images
4. Create Blog Post – Research and write blog posts with graphics
5. Create LinkedIn Post – Generate LinkedIn content with visuals
6. Create Video – Generate faceless videos with AI images, video clips, and sound effects

All outputs are logged to a Google Sheet and stored in Google Drive for easy access later.

How It Works

You send a message (voice or text) to your Telegram bot. The main agent receives it, decides which tool to use, and executes the appropriate workflow.

Example interactions:

• “Create an image for a cat food flyer with a flash sale” – Agent creates image, sends to Telegram, logs to Sheet
• “Edit that image and make it more realistic” – Agent searches database, finds image, edits it, returns result
• “Create a blog post about the effect of sleep on productivity” – Agent researches topic, writes post, creates graphic, returns everything
• “Create a video of a beaver building a house” – Agent generates 4 images, converts to videos, adds sound effects, renders final clip

The Six Tools – Deep Dive

1. Create Image Tool

Uses OpenAI’s new GPT-4o image model (the same one in ChatGPT). An image prompt agent first enhances your request into a detailed prompt, then sends it to OpenAI. The image is converted from base64, sent to Telegram, uploaded to Google Drive, and logged to your Sheet.

2. Edit Image Tool

Similar to create, but uses OpenAI’s edit endpoint. Keeps the image consistent while making requested changes. Great for iterations like “add headphones to the crocodile” or “make this more realistic.”

3. Search Image Database

An image retrieval agent that searches your Google Sheet log. Can either return an image to you (“show me that crocodile picture”) or pass the image ID to the edit tool (“edit that last image”).

4. Blog Post Tool

A blog posting agent with Tavily access for web research. Writes professional, educational content tailored to your target audience, includes sources, then generates a matching graphic using the image pipeline.

5. LinkedIn Post Tool

Same architecture as the blog tool, but with LinkedIn-specific writing style and graphic design prompts optimized for the platform.

6. Video Generation Tool

The most complex workflow. Creates 4 cohesive scene prompts, generates 4 images (via Flux/Piapi), converts each to 5-second video clips (via Runway), generates sound effects for each scene (via ElevenLabs), then renders everything together (via Creatomate). Takes 2-3 minutes per video.

Tech Stack & APIs

• n8n – Workflow automation platform (cloud or self-hosted)
• Telegram – Communication interface (voice and text)
• OpenAI GPT-4o – Image generation and editing
• OpenAI GPT-4.1 / 4.1 Mini – Text generation and reasoning (via Open Router)
• Tavily – Web research for blog/LinkedIn posts
• Piapi (Flux) – Image generation for videos (~1.5 cents/image)
• Runway – Image-to-video generation (~25 cents/5-second clip)
• ElevenLabs – Sound effect generation ($5/month starter plan)
• Creatomate – Video rendering (50 free credits on trial)
• Google Drive/Sheets – Storage and logging

Pricing Breakdown

Image Creation/Editing

• OpenAI GPT-4o Image: ~19-20 cents per image or edit
• Alternative: DALL-E 2/3 available at lower price points

Video Creation

• Flux images (Piapi): ~1.5 cents per image (4 images = 6 cents)
• Runway video: ~25 cents per 5-second clip (4 clips = $1)
• ElevenLabs sound: ~$5/month (generous credits)
• Creatomate: ~1 credit per 20-second render (50 free on trial)

Text Generation

• GPT-4.1: $2/million input tokens, $8/million output tokens
• GPT-4.1 Mini: $0.40/million input, $1.60/million output

Bottom line: A 20-second video costs roughly $1-2. A blog post with image costs about 20-25 cents. Individual images are 20 cents each.

Setup Guide

Step 1: Get the Templates

Join Nate’s free Skool community (link in video description). Search for the video title or go to YouTube Resources to find the post with all 7 JSON workflow templates, Google Sheet template, and Creatomate rendering template.

Step 2: Import Workflows

Import each of the 7 JSON files as individual workflows in n8n (the main agent + 6 tools). Use Import from File for each one.

Step 3: Link the Tools

In the main agent workflow, each tool references a sub-workflow. Make sure each tool is linked to the correct workflow you imported.

Step 4: Set Up Credentials

• OpenAI API key
• Open Router API key (for GPT-4.1)
• Google Drive/Sheets credentials
• Telegram bot token
• Tavily API key
• Piapi API key (for Flux)
• Runway API key
• ElevenLabs API key
• Creatomate API key

Step 5: Google Sheet Setup

Copy the Google Sheet template to your own Google account. This is where all outputs get logged (title, type, request, ID, link, post content).

Step 6: Creatomate Template

Create a new template in Creatomate, paste the provided script in the source editor, enable Template API Integration, and import the curl command into your n8n HTTP request node.

Key Takeaways

• One agent can manage an entire marketing content pipeline
• Sub-workflows as tools let you build modular, reusable components
• Voice or text input through Telegram makes it accessible anywhere
• Logging to Google Sheets creates a searchable content database
• Video generation requires multiple API calls with wait steps for processing
• Cost per piece of content is remarkably low (cents, not dollars)

Build Your Marketing Team

This isn’t a theoretical concept – it’s a working system you can deploy today. With free templates and a clear setup guide, you can have your own AI marketing team running in an afternoon.

The real power is in the architecture: one agent, multiple specialized tools, all working together. Once you understand this pattern, you can extend it to any workflow you can imagine.

Download the templates. Set up your APIs. Start creating content at scale.