Telegram photo to ad image and video maker

What it is

This n8n workflow turns a photo sent to a Telegram bot into two ad concepts, and gives each one an image ad and a short video ad. It runs 34 nodes in three stages. First, OpenAI reads the photo as YAML: brand, colors, fonts and what is shown. An agent named Campaign Strategy Generator then writes two concepts, which land as rows in a Google Sheet with a slogan, a character, a visual guide and three hex colors. Next, a second agent writes an image prompt for each row, Nano Banana Edit on Kie.ai redraws your photo as a 9:16 ad, and the file goes to Google Drive. Last, OpenAI looks at that image and writes a video prompt, Veo 3 Fast on Kie.ai animates it, and the clip goes to Drive too.

Who it suits

Shop owners, brand teams and social media managers who want ad drafts from a phone photo, and agencies that test a few angles on a product before anyone briefs a designer. Skip it if you do not use n8n yet, or if you want one polished ad rather than a pair of drafts that you still have to review, pick from and finish yourself.

Where to use it

Any n8n that Telegram can reach. The Telegram trigger needs a public HTTPS address, which n8n Cloud comes with and which a server of your own may need a reverse proxy or a tunnel to provide.

• n8n Cloud: import the file into a new workflow there
• Your own n8n server: the same file, with a public HTTPS address for the Telegram trigger
• Telegram: you send the photo from your phone and the bot answers in the same chat
• Google Drive and Google Sheets: the finished files and the rows live there

Setup

Get it saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Its author puts setup at about 20 minutes, which assumes the accounts exist.

1. In n8n, create a workflow, open its menu, choose Import from File and pick workflow.json.
2. Add a Telegram credential with your bot token and select it in Telegram Trigger, Send request msg and Send completed msg.
3. Open Set Bot ID and replace YOUR_BOT_TOKEN with the same token. The photo’s download link is built from it.
4. Add an OpenAI credential and select it in GPT, OpenAI Chat Model, Describe Image and Video Prompt Creator.
5. Make a sheet whose first row holds the columns below. Add a Google Sheets credential, select it in all seven Sheets nodes, and in each one replace YOUR_SHEET_ID with your sheet’s ID and pick the sheet again.
6. Make two Drive folders and add a Google Drive credential. In Upload file replace YOUR_IMAGE_FOLDER_ID, and in Upload file1 replace YOUR_VIDEO_FOLDER_ID.
7. In Create Image, Get Image, Create Video and Get Video, replace <YOUR_KIE_API_KEY> with your Kie.ai key and keep the word Bearer before it.
8. Save, press Execute workflow, send your bot a photo with a caption, then press Publish when the test works.

index, ad_copy, product, character, visual_guide, 1_color, 2_color, 3_color, image_status, image_link, video_status, video_prompt, video_aspectratio, video_link

How to use it

Send your bot a photo of a product or a character. A caption is optional. Use it to name the audience or the mood, and without one the agent falls back on a default request. The bot replies once it has read the photo. The two rows appear in the sheet a little later. The image takes at least 20 seconds and the video at least a minute. The workflow waits that long, asks Kie.ai for the result, and asks again until it reports success or failure. The bot sends a closing message for each concept whose video is in Drive.

Example

A photo of a canned iced tea, sent with the caption ads for office workers who skip lunch. OpenAI describes the can, and the strategist returns two concepts in this shape, trimmed here.

{
  "index": "01",
  "ad_copy": "…",
  "product": "…",
  "character": "…",
  "visual_guide": "…",
  "1_color": "#…",
  "2_color": "#…",
  "3_color": "#…",
  "image_status": "create",
  "video_status": "create"
}

The sheet then shows create turning to done for the image and again for the video, with a Drive link in image_link and video_link. A status of error means Kie.ai reported a failed task.

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. A credential is a key n8n keeps for you and hands to the node you pick it in. Kie.ai makes the image and the clip on its own servers, which takes a while, so the workflow keeps waiting and asking. Run it once with a test photo. Then open each node: it shows what it received and what it sent, which is where a wrong key or sheet ID turns up.

For experts. Four AI steps do the thinking and two Kie.ai models do the drawing.

Two habits of the sheet logic matter if you reuse it. Each run numbers its concepts 01 and 02, and the match column is index. A new photo therefore replaces the rows of the last one, while the files in Drive stay, so rows you want to keep need copying out first. The node Get Rows to Create reads every row whose image_status is create, so a row left behind by a stopped run is picked up by the next one. One loose end: the image agent reads a text_watermark field that nothing in the file fills.

Good to know

Each run costs money. The open bot and the photo’s link are the risks to plan for.

• Billing: OpenAI and Kie.ai each bill their own part, and a run makes several calls to each
• Open bot: the trigger has no filter on who writes, so anyone who finds your bot can send it a photo and spend your credits. Add an IF node after it that checks the sender’s ID
• Token in the link: the photo reaches OpenAI and Kie.ai as a Telegram file link, and that link contains your bot token. Use a bot made for this workflow
• No photo, no run: a message without a photo stops at Get Image Path with an error
• No cap on retries: the status checks repeat until Kie.ai answers, so watch your first runs
• Credentials: the JSON holds none, and nothing from its author’s account