Create and publish AI videos from Telegram with VEED and Blotato

Create & Publish AI Videos from Telegram Chat with VEED and Blotato

Overview

This n8n workflow creates a conversational AI video agent accessible through Telegram. Users chat with a bot to request AI-generated talking-head videos using VEED's MCP tools, then optionally publish to 9 social media platforms via Blotato with a single approval tap.

Output: AI-generated talking-head videos delivered in Telegram and published to TikTok, YouTube, Instagram, LinkedIn, Facebook, X, Threads, Bluesky, and Pinterest.

---

What It Does

User Message (Telegram) → AI Agent → VEED MCP tools create video → Response sent via Telegram
 ↕ Memory ↓ (if video URL detected)
 Approval prompt (Approve/Reject)
 ↓ Approved
 Upload to Blotato → Publish to 9 platforms in parallel
 ↓
 Merge results → Summary sent via Telegram

Flow Breakdown

Step	Component	What Happens
1. Message Received	📩 Telegram Trigger	User sends a text message to the bot
2. AI Processing	🤖 AI Video Agent (gpt-5-nano)	Interprets request, decides which VEED tools to use
3. Context Retention	💾 Conversation Memory	Maintains conversation history per chat (20-message window)
4. Video Creation	🎬 VEED MCP Tools	Lists characters/voices, confirms details, creates video via VEED Fabric
5. Response Delivery	📤 Send Agent Response	Sends agent text back to user in Telegram
6. URL Extraction	📋 Extract Video URL	Detects the VIDEO_READY: marker in agent output
7. Approval Prompt	📩 Publish to Social Media?	Telegram sendAndWait with Approve/Reject buttons (24h timeout)
8. Upload	📤 Upload Video to Blotato	Uploads video file to Blotato CDN, returns hosted URL
9. Publish	TikTok, YouTube, Instagram, LinkedIn, Facebook, X, Threads, Bluesky, Pinterest	All 9 platform nodes post in parallel using the Blotato URL
10. Summary	📤 Send Publish Summary	Confirms publication back to user in Telegram

---

Required Connections

n8n Credentials

Node	Credential Type	Where to Get
📩 Telegram Trigger	Telegram Bot API	Create bot via @BotFather in Telegram
🧠 OpenAI Chat Model	OpenAI API Key	https://platform.openai.com/api-keys
🎬 VEED MCP Tools	MCP OAuth2	Your VEED OAuth2 client credentials
📤 Send Agent Response	Telegram Bot API	Same bot token as trigger
📩 Publish to Social Media?	Telegram Bot API	Same bot token as trigger
📤 Upload Video to Blotato	Blotato API	https://my.blotato.com/settings → API
TikTok / YouTube / Instagram / ...	Blotato API	Same Blotato credential as upload

Community Node Required

Install @blotato/n8n-nodes-blotato via Settings → Community Nodes → Install in n8n.

---

Configuration Options

AI Agent System Prompt

The AI Agent's behavior is defined by its system prompt. Edit the 🤖 AI Video Agent node to customize:

You are a friendly AI video creation assistant powered by VEED.
You help users create professional talking-head videos with AI avatars through Telegram.

Video Creation Flow:
1. Understand their need - Ask what video they want
2. Choose a character - Show available AI avatars
3. Choose a voice - Ask preferred language/locale
4. Confirm details - Summarize before creating
5. Create the video - Generate with confirmed parameters
6. Track progress - Poll generation status
7. Deliver result - Share the video URL with VIDEO_READY: marker

Agent Configuration

Field	Default	Description
Model	gpt-5-nano	OpenAI model used for reasoning
Max Iterations	15	Maximum tool-calling rounds per message
Memory Window	20 messages	How many past messages the agent remembers
Session Key	{{ $json.message.chat.id }}	Isolates memory per Telegram chat

Available VEED MCP Tools

Tool	Required Parameters	Description
list_workspaces	—	Show available workspaces
list_characters	— (optional: gender)	Browse AI avatar characters
list_voices	locale (e.g. "en")	Browse voices (optional: gender filter)
confirm_fabric_video	script, voiceId, characterId	Preview video details before creating
create_fabric_video	script, voiceId, characterId	Create a talking-head video (optional: aspectRatio)
get_generation_status	jobId	Check video generation progress
get_credit_balance	—	Check remaining VEED credits

Blotato Platform Configuration

Each platform node needs its accountId configured (select from dropdown after connecting Blotato credentials):

Platform	Extra Configuration
TikTok	—
YouTube	Title, privacy status (private by default), notify subscribers
Instagram	—
LinkedIn	—
Facebook	Page ID (select from dropdown)
Twitter / X	—
Threads	—
Bluesky	—
Pinterest	Board ID

Disable any platform node you don't use.

Caption Configuration

Edit the ⚙️ Set Caption node to customize what gets posted:

• caption: defaults to first 200 characters of the agent's response
• video_url: automatically extracted from the VIDEO_READY: marker

---

Conversation Examples

User Says	Agent Does
"Hi, what can you do?"	Introduces itself and explains video creation capabilities
"Create a video about AI trends"	Asks about character/voice preferences, then guides through creation flow
"Show me available characters"	Calls list_characters, formats results for easy selection
"Use a female English voice"	Calls list_voices with locale "en" and gender filter, shows options
"Yes, create it!"	Calls create_fabric_video, then immediately checks get_generation_status
"Is my video ready?"	Calls get_generation_status with the last jobId from memory
Taps "Approve"	Video uploads to Blotato, publishes to all 9 platforms, summary sent
Taps "Reject"	"No problem! Your video won't be published to social media."

---

Output

Per Video Generated

Asset	Format	Delivered Via
Video URL	VEED hosted MP4 link	Telegram message
Approval prompt	Approve/Reject buttons	Telegram sendAndWait
Social media posts	Video + caption	TikTok, YouTube, Instagram, LinkedIn, Facebook, X, Threads, Bluesky, Pinterest
Publish summary	Text confirmation	Telegram message

---

Estimated Costs Per Video

Service	Usage	Approximate Cost
OpenAI gpt-5-nano	~1-3K tokens per conversation turn	~$0.005-0.02
VEED Fabric	1 video render (~8 credits/sec)	~$0.10-0.20
Blotato	Media upload + 9 platform posts	Included in paid plan
Telegram Bot API	Messages	Free
Total		~$0.10-0.25 per video

> Costs vary based on conversation length, video duration, and current API pricing.

---

Setup Checklist

Step 1: Import Workflow

• Import create-and-publish-ai-videos-from-telegram-chat-with-veed-and-blotato.json into n8n

Step 2: Install Blotato Community Node

• Go to Settings → Community Nodes → Install
• Enter @blotato/n8n-nodes-blotato and install

Step 3: Create Telegram Bot

• Open Telegram and message @BotFather
• Send /newbot and follow the prompts to create your bot
• Copy the bot token provided by BotFather

Step 4: Configure n8n Credentials

• Click on 📩 Telegram Trigger node → Add Telegram Bot credential with your token
• Click on 🧠 OpenAI Chat Model node → Add OpenAI API Key credential
• Click on 🎬 VEED MCP Tools node → Add MCP OAuth2 credential with your VEED OAuth2 client credentials
• Click on 📤 Send Agent Response node → Select the same Telegram Bot credential
• Click on 📩 Publish to Social Media? node → Select the same Telegram Bot credential
• Click on 📤 Upload Video to Blotato node → Add Blotato API credential (from blotato.com settings)

Step 5: Configure Blotato Platforms

• Open each platform node (TikTok, YouTube, Instagram, etc.)
• Select your connected account from the accountId dropdown
• For YouTube: set title and privacy preferences
• For Facebook: select your Page from the dropdown
• For Pinterest: set Board ID
• Disable any platform nodes you don't need

Step 6: Activate & Test

• Activate the workflow by clicking Publish and toggling it active
• Send a test message to your bot in Telegram (e.g., "Hi, what can you do?")
• Verify the bot responds with its capabilities
• Create a video and test the full flow including the approval prompt
• Approve publishing and verify posts appear on your social accounts

---

Limitations & Notes

Technical Limitations

• Telegram activation: The workflow must be published and active — simply executing it is not enough for the Telegram trigger to receive messages.
• Memory window: Only the last 20 messages are retained per conversation. Older context is lost.
• VEED processing: Video generation takes 1-5 minutes. The agent reports status but the user must wait.
• Concurrent users: Multiple users are supported — each Telegram chat gets independent conversation memory keyed by chat.id.
• Approval timeout: The publish prompt expires after 24 hours if not answered.
• VIDEO_READY marker: The agent must include VIDEO_READY: <url> in its response for the publishing flow to trigger. This is enforced via the system prompt.
• Blotato paid plan: API access requires a paid Blotato subscription.

Content Considerations

• The AI agent autonomously decides which VEED tools to use based on the user's request
• Users can request specific characters, voices, and scripts through natural conversation
• The agent cannot edit or modify videos after creation
• Video output depends on available VEED characters and voices
• Each second of generated video costs approximately 8 VEED credits
• Published captions default to the first 200 characters of the agent's response — edit the ⚙️ Set Caption node to customize

Best Practices

1. Set bot commands: Use @BotFather's /setcommands to add help commands for your users
2. Test with simple requests first: Start with "list available characters" before attempting full video creation
3. Monitor token usage: Longer conversations consume more OpenAI tokens per turn
4. Customize the system prompt: Tailor the agent's persona, default locale, and guidelines to your use case
5. Disable unused platforms: Disable platform nodes you haven't connected to avoid errors
6. YouTube defaults to private: Videos are published as private by default — change in the YouTube node if needed

---

Troubleshooting

Issue	Solution
Bot not responding	Ensure workflow is published and active, not just executed manually
"Unauthorized" from Telegram	Verify bot token is correct in n8n Telegram credential
VEED MCP connection failed	Re-authorize the VEED OAuth2 credential in n8n
Agent doesn't use VEED tools	Verify 🎬 VEED MCP Tools node is connected to the AI Agent node
Memory not working across messages	Confirm session key expression {{ $('📩 Telegram Trigger').item.json.message.chat.id }} is correct
Video generation timeout	VEED processing can take up to 5 minutes — agent should poll get_generation_status
"Model not found" error	Verify OpenAI API key has access to gpt-5-nano model
No approval prompt after video	Check the agent includes VIDEO_READY: marker — verify system prompt is intact
Blotato upload fails	Verify Blotato API credential is valid and plan is active
Platform publish fails	Ensure accountId is configured in the failing platform node
"Can't parse entities" error	The Send Agent Response node should have no parse_mode set (plain text)
Blotato node not found	Install @blotato/n8n-nodes-blotato via Settings → Community Nodes

---

Version History

Version	Date	Changes
1.0	Mar 2026	Initial release with Telegram bot, AI Agent, VEED MCP integration, conversation memory
2.0	Apr 2026	Added Blotato multiplatform publishing (9 platforms), Telegram approval flow, video URL extraction, caption configuration

---

Credits

Built with:

• n8n - Workflow automation
• OpenAI - AI agent reasoning (gpt-5-nano)
• VEED - AI video generation (via MCP)
• Blotato - Multiplatform social media publishing
• Telegram - Chat interface & approval flow
• LangChain - Agent framework and memory