Influencer Content Creation is an AI content creation system for social networks based on Telegram management

Process description

This automation is an intelligent content creation system for influencers that works through the Telegram interface. The system uses a central AI agent to manage the process, specialized agents to generate prompts, creates videos via Fal.ai (Google Veo3) or edits photos through Google Gemini, generates captions for social networks, and automatically publishes content to Instagram via the Blotato API.

API keys and services:

1. Telegram Bot API - for the user interface
2. OpenAI API Key - for all AI agents (gpt-4.1-mini)
3. Fal.ai API - to generate videos via Google Veo3
4. Google Gemini API - for editing images
5. ImgBB API - for image hosting
6. Blotato API - for posting on Instagram
7. Google Drive OAuth2 - to access the original image

System architecture by blocks

SECTION 1: RECEIVING AND PROCESSING REQUESTS

1.1 Telegram Trigger — Entry Point

Purpose: Tracks all incoming messages from users Settings:

• Updates: message
• Additional Fields: standard settings
• Credentials: your credentials from the telegram bot

1.2 Central Agent - Central Coordinator

Purpose: An intelligent agent that understands user requests and coordinates the work of other agents

System prompt:

You're the central management agent for the SMM system.

Based on the user's request “Upload a video” or “Post a photo”, use your Photo Agent and Video Agent tools and be sure to inform the user that the task is in progress through the telegram tool and send the result of the Photo Agent or Video Agent tool in English further down the workflow.

IMPORTANT! Send only the output of your instruments as a result, don't add any comments from yourself!

Connected components:

• OpenAI Chat Model (gpt-4.1-mini) - main language model
• Simple Memory - memory
• Photo Agent - a tool for generating photo prompts
• Video Agent - a tool for generating prompts for videos
• Send a text message in Telegram - user notification tool

1.3 Switch - Content Router

Purpose: Sends a request to the appropriate processing branch based on Central Agent output Routing conditions:

• Output 0 “Video”: $ ('Telegram Trigger') .item.json.message.text equals “video”
• Output 1 “Photo”: $ ('Telegram Trigger') .item.json.message.text contains “photo”

SECTION 2: SPECIALIZED AI AGENTS

2.1 Photo Agent - Photo Prompt Generator

Purpose: Generates detailed prompts to generate static influencer images

System prompt (full structure):

Prompt for Nano-banana Prompt Generation Agent AI for TikTok photos

‍

Main role:

You are a specialist in creating incentives for Nano-banana. Your task is to generate one unique detailed prompt for creating a photo of a female blogger for TikTok content.

‍

Base character:

Young woman, 20-25 years old, long wavy blond hair, Caucasian appearance, charismatic, natural beauty, expressive facial expressions.

‍

PROMPT STRUCTURE (use as a single text):

A young blonde woman with long wavy hair, [facial expression/emotion], [posture/body position], wearing [detailed clothing description], [accessories if any], in/at [location details], [lighting], [shooting style], [image quality], vertical portrait orientation, Instagram/TikTok style photo

Randomized items:

• Emotions/facial expressions (12+ options)
• Poses (13+ options)
• Clothing (10+ detail sets)
• Accessories (10+ types)
• Locations (12+ detailed locations)
• Lighting (9+ options)
• Shooting styles (9+ options)

Connected components:

• OpenAI Chat Model1 (gpt-4.1-mini)
• Simple Memory

2.2 Video Agent - Video Prompt Generator

Purpose: Creates detailed prompts for generating dynamic video content

System prompt (full structure):

Prompt for Veo3 Prompt Generation Agent AI for TikTok videos

‍

Main role:

You're a Google Veo3 Prompt Developer. Your task is to generate one unique detailed prompt to create a short TikTok video (5-15 seconds) with a female blogger.

‍

PROMPT STRUCTURE (use as a single text):

A young blonde woman with long wavy hair [emotion/condition], wearing [clothes], in/at [location]. [Shooting type] shot, [camera movement if any]. She [actions, gestures, facial expressions] while saying: “[speech up to 30 words]”. [Lighting], [atmosphere], [additional background details]. Duration: [5-15] seconds, vertical format 9:16, TikTok style, [shooting style].

Thematic categories:

• 14 different topics (morning routines, relationships, studies, working from home, etc.)
• 6 video formats (reaction, life hack, POV, story, opinion, answer)
• 11 emotional states
• 10 clothing options
• 10 locations
• 8 visual styles

Connected components:

• OpenAI Chat Model2 (gpt-4.1-mini)
• Simple Memory

SECTION 3: VIDEO PROCESSING BRANCH (Upper Stream)

3.1 Create video — Initiate generation

Purpose: Sends a video creation request to the Fal.ai API HTTP Request settings:

• Method: POST
• URL: https://queue.fal.run/fal-ai/veo3/fast/image-to-video
• Headers: Content-Type: application/json
• Authentication: HttpHeaderAuth (Fal credentials)
• Body:

json

{

“prompt”: “{{$json.output.replace (/\” /g, '\\\ "') .replace (/\\n/g, '')}}”,

“image_url”: "https://i.ibb.co/XkkhHXh8/Influencer.png”

}

3.2 Wait — Waiting for processing

Purpose: Pause for video processing on the Fal.ai side Settings:

• Amount: 30
• Type: seconds

3.3 Verification — Status Monitoring

Purpose: Checks video readiness via API Settings:

• URL: https://queue.fal.run/fal-ai/veo3/requests/ {{$ ('Create video') .item.json.request_id}} /status
• Authentication: HttpHeaderAuth (Fal credentials)

3.4 If - Conditional Logic

Purpose: Checks video generation status Condition:

• Condition: {{$json.status}} equals “COMPLETED”
• True Output: Go to “Get video”
• False Output: Return to “Wait” to check again

3.5 Get video - Download result

Purpose: Gets the finished video from Fal.ai Settings:

• URL: https://queue.fal.run/fal-ai/veo3/requests/ {{$ ('Create video') .item.json.request_id}}
• Authentication: HttpHeaderAuth (Fal credentials)

3.6 Send a text message2 - Notification of the finished video

Purpose: Sends the user a link to the finished video Settings:

• Chat ID: {{$ ('Telegram Trigger') .item.json.message.chat.id}}
• Text: Your video is ready: <a href="{{ $json.video.url }}">Click here</a>
• Parse Mode: HTML

3.7 Captions — Video Caption Generator

Purpose: Creates an attractive caption for video content System prompt:

Your task is to generate a short title in English for an Instagram post based on the received prompt.

Connected components:

• OpenAI Chat Model3 (gpt-4.1-mini)
• Simple Memory

3.8 Set Accounts - Configure video accounts

Purpose: Prepares data for video publishing JSON structure:

json

{

“blotato_api_key”: “”,

“instagram_id”: “”,

“youtube_id”: “”,

“tiktok_id”: “”,

“facebook_id”: “”,

“facebook_page_id”: “”,

“threads_id”: “”,

“twitter_id”: “”,

“linkedin_id”: “”,

“pinterest_id”: “”,

“pinterest_board_id”: “”,

“bluesky_id”: “”,

“video_caption”: “{{$json.output}}”

}

Insert the ID you received after linking to the social network you need

3.9 Upload — Uploading videos to Blotato

Purpose: Uploads a video file to Blotato for further publishing Settings:

• Method: POST
• URL: https://backend.blotato.com/v2/media
• Headers: blotato-api-key: {{$ ('Set Accounts') .item.json.blotato_api_key}}
• Body: url: {{$ ('Get video') .item.json.video.url}}

3.10 INSTA — Posting videos to Instagram

Purpose: Creates an Instagram video post Settings:

• URL: https://backend.blotato.com/v2/posts
• Body:

json

{

“post”: {

“target”: {

“targetType”: “instagram”,

“isYourBrand”: false,

“DisabledDuet”: false,

“PrivacyLevel”: “PUBLIC_TO_EVERYONE”,

“IsaiGenerated”: false,

“DisabledStitch”: false,

“DisabledComments”: false,

“isBrandedContent”: false

},

“content”: {

“text”: “{{$ ('Set Accounts') .item.json.video_caption}}”,

“platform”: “instagram”,

“MediaURLS”: ["{{$json.url}}"]

},

“accountID”: “{{$ ('Set Accounts') .item.json.instagram_id}}”

}

}

SECTION 4: PHOTO PROCESSING BRANCH (Lower stream)

4.1 Download file - Download the base image

Purpose: Downloads the influencer's original image from Google Drive Settings:

• Operation: download
• File ID: your ID file (Influencer.png)
• Credentials: your credentials

4.2 Extract from File - Convert to Base64

Purpose: Converts a binary file to a Base64 string for the API Settings:

• Operation: BinaryToProperty
• Options: standard

4.3 generate_image - AI photo editing

Purpose: Edits an image via Google Gemini API Settings:

• Method: POST
• URL: https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image-preview:generateContent
• Authentication: HttpHeaderAuth (Gemini credentials)
• Body:

json

{

“model”: “gemini-2.5-flash-image-preview”,

“contents”: [

{

“parts”: [

{

“text”: “Edit the photo by this instruction: {{$ ('Central Agent') .item.json.output}}. The edited image must be in high quality! And do not make any comments just send the edited photo.”

},

{

“inlineData”: {

“mimeType”: “image/jpeg”,

“data”: “{{$json.data}}”

}

}

]

}

]

}

4.4 Convert to File - Reverse conversion

Purpose: Converts a Base64 response back to a file Settings:

• Operation: Tobinary
• Source Property: candidates [0] .content.parts [0] .inlineData.data

4.5 Get URL - Uploading images to hosting

Purpose: Uploads the processed photo to ImgBB to get a public URL Settings:

• Method: POST
• URL: https://api.imgbb.com/1/upload
• Authentication: your credentials
• Content Type: multipart/form-data
• Body: image (FormbinaryData): data

4.6 Send a text message1 - Notification about the finished photo

Purpose: Sends the user a link to the finished photo Settings:

• Chat ID: {{$ ('Telegram Trigger') .item.json.message.chat.id}}
• Text: Your photo is ready {{$json.data.url}}

4.7 Captions1 - Photo Caption Generator

Purpose: Creates an attractive caption for photo content System prompt: Similar to Captions for videos Connected components:

• OpenAI Chat Model4 (gpt-4.1-mini)

4.8 Set Accounts1 - Configure photo accounts

Purpose: Prepares data for publishing photos Structure: Identical to Set Accounts but uses Captions1's output

4.9 Upload1 - Uploading photos to Blotato

Purpose: Uploads photos to Blotato for publication Settings:

• URL: https://backend.blotato.com/v2/media
• Headers: blotato-api-key: {{$ ('Set Accounts1') .item.json.blotato_api_key}}
• Body: url: {{$ ('Get URL') .item.json.data.url}}

4.10 INSTA1 — Posting a photo to Instagram

Purpose: Creates a post with a photo on Instagram Settings: Identical to INSTA but uses data from Set Accounts1

SECTION 5: AUXILIARY COMPONENTS

5.1 Send a text message in Telegram (Tool)

Purpose: A tool for Central Agent to send notifications Settings:

• Chat ID: {{$ ('Telegram Trigger') .item.json.message.chat.id}}
• Text: “The task is at work”

5.2 Memory components

All Simple Memory nodes use:

• Session ID Type: CustomKey
• Context Window Length: 10
• Unique Session Keys for isolating contexts

Node connection diagram

Main stream:

Telegram Trigger → Central Agent → Switch

Video thread (upper stream):

Switch (Output 0: Video) → Create Video → Wait → Check → If

├─ If (True: COMPLETED) → Get Video → Send a text message2 → Captions → Set Accounts → Upload → INSTA

└─ If (False: NOT COMPLETED) → Wait (cycle)

Photo branch (lower stream):

Switch (Output 1: Photo) → Download file → Extract from File → generate_image → Convert to File → Get URL → Send a text message1 → Captions1 → Set Accounts1 → Upload1 → INSTA1

Connecting AI components to Central Agent:

OpenAI Chat Model ─

Simple Memory ──────→ Central Agent

Photo Agent ───────

Video Agent ───────

Send a text msg ──

Connections to specialized agents:

OpenAI Chat Model1+ Simple Memory1 → Photo Agent

OpenAI Chat Model2+ Simple Memory2 → Video Agent

OpenAI Chat Model3+ Simple Memory3 → Captions

OpenAI Chat Model4+ Simple Memory4 → Captions1

Required services and their settings

Telegram Bot setup:

1. Create a bot via @BotFather
2. Get a token (API key)
3. Install a webhook for the trigger

OpenAI setup:

1. Get the OpenAI API key
2. Create “OpenAI account 6" credentials
3. Choose the gpt-4.1-mini model for all agents
4. Set up limits and billing

Setting up Fal.ai:

1. Sign up for fal.ai
2. Get an API key
3. Create HTTP Header Auth credentials “Fal”
4. Activate access to the Veo3 model

Google Gemini setup:

1. Get the Google AI Studio API key
2. Create “Gemini” HTTP Header Auth credentials
3. Activate gemini-2.5-flash-image-preview

Setting up Google Drive:

1. Set up OAuth2 for Google Drive
2. Download the Influencer.png base image
3. Select by File ID:

ImgBB setup:

1. Sign up to imgbb.com
2. Get an API key
3. Create Query Auth credentials

Blotato setup:

1. Sign up to blotato.com
2. Connect an Instagram Business account
3. Get an API key and Instagram ID
4. Set up publishing permissions

System capabilities

Content generation:

• Video content: 5-15 second TikTok videos with AI animation
• Photo content: Stylized photos for Instagram
• Unique offers: Each request generates new unique content
• Signatures: Automatic generation of engaging signatures

AI functionality:

• Multi-agent architecture: Specialized agents for different tasks
• Contextual memory: Personalization for every user
• Natural language: Understanding commands in Russian
• Intelligent routing: Automatic content type selection

Automation:

• Full cycle: From a team to publishing without human intervention
• Multiplatform: Ready to publish on multiple platforms
• Cloud processing: All processes in the cloud
• Scalability: Parallel processing of requests

Use cases

Scenario 1: Creating and publishing videos

👤 User: video

🤖 Central Agent: Task at work

🤖 Video Agent: [generates a unique prompt]

⚙️ System: [creates videos via Veo3]

📹 Result: 10-second video reacting to the trend

🤖 Bot: Your video is ready: [link]

📝 Captions: [generates a signature]

📱 Instagram: [automated post]

Scenario 2: Creating and publishing photos

👤 User: post a photo

🤖 Central Agent: Task at work

🤖 Photo Agent: [generates a unique prompt]

⚙️ System: [edits photos via Gemini]

📸 Result: Stylized photo in a cafe

🤖 Bot: Your photo is ready [link]

📝 Captions1: [generates a signature]

📱 Instagram: [automated post]

Scenario 3: Publication Series

👤 User: video

[in 2 minutes]

👤 User: photo

[in 3 minutes]

👤 User: more videos

‍

The result: 3 unique posts with different content thanks to agent randomization

System application

For influencers:

• Regular content: Daily posts without filming
• Variety: Automatic rotation of themes and styles
• Consistency: The character's unified visual style
• Save time: 2 minutes instead of hours to create a post

For SMM agencies:

• Scaling: Manage dozens of accounts
• Automation: Fully autonomous operation
• Creativity: AI generation of unique ideas
• Analytics: Tracking via Blotato

For brands:

• Virtual ambassadors: Digital brand representatives
• A/B testing: Quick validation of creative concepts
• Trendness: Prompt response to viral topics
• ROI: 90% reduction in content costs

The result of the system

Output data:

• Video: 5-15 second 9:16 vertical videos
• Photo: High-quality images for Instagram
• Signatures: SEO-optimized texts in English
• Publication: Automated Instagram posts
• Notifications: Links to ready-made content on Telegram

Performance metrics:

• Video generation time: 60-120 seconds
• Photo generation time: 30-45 seconds
• Content quality: HD/4K resolution
• Uniqueness: 100% original content
• Publication success: 99.9% uptime

System advantages:

• Speed: 50 times faster than traditional production
• Price: 20 times cheaper than hiring a content team
• Accessibility: Works 24/7 seven days a week
• Scale: Unlimited content
• Quality: Professional level of AI generation

ROI and economy:

• Savings on production
• Increased coverage
• Decrease in CPA
• Payback
• Scaling: Linear growth without increasing costs

This system is a full-fledged AI content factory that turns simple text commands into professional social media content with automatic publication!