Daily AI news avatar video for social

What it is

This n8n workflow makes one short news video a day and posts it. At 10 in the morning an AI agent reads Hacker News, picks the day’s AI story most likely to spread, opens the article and its comments, and writes a 30-second monologue. Two more model calls write a long and a short caption. HeyGen turns the script into a vertical avatar video, the workflow waits eight minutes, and Blotato uploads the finished file and posts it to Instagram, YouTube, TikTok and X. Five more platforms are wired in but switched off.

1. Each morning the agent picks and reads an AI story
2. It writes a 30-second script and two captions
3. HeyGen renders a vertical avatar video
4. Eight minutes later the video is fetched
5. Blotato uploads it and posts to four platforms

Who it suits

Creators and small media brands who want a news clip every day without filming. Skip it if every post needs a human check. Nothing in the file pauses for approval, and the four posts it makes go out public.

Where to use it

Any n8n that can reach OpenAI, Hacker News, HeyGen and Blotato will run it. The posts go to the accounts you have connected in Blotato. Pick only the platforms you use.

• n8n Cloud or your own server: import the file, and publish it so the daily schedule runs
• Instagram, YouTube, TikTok and X: the four platforms the file posts to as it comes
• Facebook, LinkedIn, Threads, Pinterest and Bluesky: wired in but switched off, so you opt in per platform
• A review step: add a node before Upload to Blotato if you want a person to approve each video

Setup

Get it saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Its author puts setup at about 10 minutes, and that leaves out creating the avatar and connecting the social accounts.

1. In n8n, create a workflow, open its menu, choose Import from File and pick workflow.json.
2. Add an OpenAI credential and select it in Write Script, Write Long Caption and Write Short Caption.
3. In Setup Heygen, replace YOUR_HEYGEN_API_KEY, YOUR_AVATAR_ID and YOUR_VOICE_ID. Replace YOUR_BACKGROUND_VIDEO_URL with the address of a clip that loops behind the avatar.
4. In Prepare for Publish, replace YOUR_BLOTATO_API_KEY and the account IDs of the platforms you post to: YOUR_INSTAGRAM_ID, YOUR_YOUTUBE_ID, YOUR_TIKTOK_ID and YOUR_X_ID.
5. In Write Long Caption, put your own tag where the hashtag line says #yourbrand.
6. Change the two titles the file posts as written: TEST VIDEO in the YouTube node and n8n TEST AVATAR in Create Avatar Video.
7. Check the hour in Schedule Trigger and the time zone n8n uses.
8. Switch off every platform but one and press Execute workflow once. Look at the post. Then switch the rest back on and press Publish.

How to use it

Nothing to press once it is published. Each morning it runs the whole chain. The clip appears on the platforms that are switched on. Open the run under Executions to see the script, the captions and the video’s address.

Example

One morning the front page holds a story about a new open model. The agent opens the article and the comments and writes a script of about 30 seconds, with two hooks first and the article’s details from the third sentence on. The two captions follow, HeyGen renders a clip of 720 by 1280, and eight minutes later Blotato posts it. Prepare for Publish holds the captions in this shape:

{
  "final_text_long": "<a summary paragraph, three questions, hashtags>",
  "final_text_short": "<two sentences>"
}

For newcomers and experts

If you are new. n8n draws a workflow as boxes joined by lines, and each box, a node, does one job. HeyGen makes the avatar video from a script. Blotato is a service with one posting API for many social accounts, so n8n needs one key instead of nine logins. The five sticky notes on the canvas are only labels and reminders. Run it by hand first. Every node then shows what it received and sent.

For experts. Three models and two services share the work, and the posting nodes all use one request shape.

The long caption goes to Instagram, YouTube and TikTok, and to Facebook and LinkedIn when they are on. The short one goes to X, and to Threads, Bluesky and Pinterest when they are on. The Pinterest branch also makes an image with the OpenAI node from the long caption, since the file’s own note says Blotato’s API could not take a video for Pinterest or Bluesky.

Good to know

It posts in public with nobody checking the script, and it spends money on four services.

• Billing: OpenAI bills the script and the captions, HeyGen and Blotato bill on their own plans, and the file’s note says HeyGen’s free plan is not enough
• No approval: the four enabled nodes post as soon as the upload finishes, and the YouTube post is public and notifies subscribers
• Unchecked facts: the script is the model’s reading of one article, and no step checks its figures before it goes out
• AI label on TikTok: the TikTok body sets isAiGenerated to false, though the video is an avatar. Set it to true to label it
• One fixed wait: Wait pauses eight minutes with no status loop, so a render that takes longer leaves Get Avatar Video without a video address and the upload fails. Raise it for long scripts
• Switched-off branches: Bluesky posts text only, and the Pinterest body has a stray comma after its text field, so remove one before you switch that node on
• Where data goes: the article and comments go to OpenAI, the script to HeyGen, and the video and captions to Blotato and then to each platform
• Credentials: the JSON holds none, and nothing from its author’s accounts or channels