Auto-generate YouTube SEO content & thumbnail with GPT-4o

Turn video scripts from Google Sheets into English, generate titles/tags/descriptions/thumbnail prompts, and call an image API to produce thumbnails, all written back to the sheet.

Who

YouTube creators who need to batch-produce SEO assets

You get

When a new row is added, automatically produce English script, 10 candidate titles, tags, keywords, SEO description, thumbnail AI prompt, thumbnail tagline, and the final thumbnail image URL

You need

n8n · google-sheets

Steps

1. Connect Google Sheets trigger

   Add a Google Sheets Trigger node in n8n, set event to rowAdded, and bind your spreadsheet and sheet. The workflow triggers when a new row is added.

2. Check if translation to English is needed

   Add an If node with conditions: Roman Hindi Script is not empty AND English script is empty. If true, proceed to translation; otherwise go to No Operation and do nothing.

3. Translate script to English with GPT-4o

   Add an OpenAI node, select model gpt-4o. Pass the Roman Hindi Script field and use the prompt below to request English translation.

   Execution prompt · Source text。OpenAI node messages.values[0].content

   Copy prompt

   ={{ $json['Roman Hindi Script'] }}
   
   Convert this hindi roman script in to english script.

4. Write English script back to Google Sheets

   First use a Set node to extract message.content, then use a Google Sheets node to update the same row, writing back Number, Roman Hindi Script, and English script columns, using Number as the matching key.

5. Generate SEO content and thumbnail prompt

   Add an OpenAI node, select model chatgpt-4o-latest, pass the English script, and use the prompt below to generate 10 titles, tags, keywords, SEO description, thumbnail AI prompt, and thumbnail tagline in one call.

   Execution prompt · Source text。OpenAI node messages.values[0].content (Sticky Note1)

   Copy prompt

   ={{ $json['English script'] }}
   
   Use the provided English script and generate:
   Generate the following:
   
   1️⃣ 10 Best Engaging YouTube Titles (≤ 70 characters each)
   
   2️⃣ Best Engaging YouTube Tags (comma-separated, up to 20)
   
   3️⃣ Best Engaging YouTube Keywords (comma-separated, up to 20)
   
   4️⃣ SEO-Friendly YouTube Video Description (optimized for 2025 algorithm, includes keywords, tags, hook, timestamps, engagement call-to-action, hashtags, and Google Search Keywords section)
   
   5️⃣ Thumbnail AI Prompt (realistic, describes main idea, bold clickable text, color scheme, visual elements, emotion, layout, realistic photography style, 1280×720, 16:9, suitable for MidJourney/DALL·E/Runware)
   
   6️⃣ Thumbnail Tagline (single catchy 3–5 word tagline)

6. Parse structured output with JS code

   Add a Code node (JavaScript) and run the code below to extract the 6 sections from the OpenAI response and output structured JSON.

   Execution prompt · Source text。Code node jsCode field (Sticky Note2)

   Copy prompt

   const raw = $input.first().json.message.content;
   
   // Helper function to extract section between markers
   function extractSection(regex, input) {
     const match = input.match(regex);
     return match ? match[1].trim() : "";
   }
   
   // 1️⃣ Titles
   let titlesRaw = extractSection(/1️⃣.*?\n([\s\S]*?)2️⃣/, raw);
   let titles = titlesRaw
     .split(/\n|-/)
     .map(t => t.replace(/^\d+\.?\s*/, "").trim())
     .filter(t => t.length > 0);
   
   // 2️⃣ Tags
   let tags = extractSection(/2️⃣.*?\n([\s\S]*?)3️⃣/, raw).replace(/\n/g, "").trim();
   
   // 3️⃣ Keywords
   let keywords = extractSection(/3️⃣.*?\n([\s\S]*?)4️⃣/, raw).replace(/\n/g, "").trim();
   
   // 4️⃣ SEO Description
   let description = extractSection(/4️⃣.*?\n([\s\S]*?)5️⃣/, raw).trim();
   
   // 5️⃣ Thumbnail AI Prompt
   let thumbnailPrompt = extractSection(/5️⃣.*?\n([\s\S]*?)6️⃣/, raw).trim();
   
   // 6️⃣ Thumbnail Tagline
   let tagline = extractSection(/6️⃣.*?\n([\s\S]*)$/, raw).replace(/["\n]/g, "").trim();
   
   return [
     {
       json: {
         "YouTube Titles": titles,
         "YouTube Tags": tags,
         "YouTube Keywords": keywords,
         "SEO Description": description,
         "Thumbnail AI Prompt": thumbnailPrompt,
         "Thumbnail Tagline": tagline
       }
     }
   ];

7. Write SEO data back to Google Sheets

   Use a Google Sheets node to update the same row, writing back Youtube Title, Youtube Tags, Youtube Keywords, Youtube Description, Youtube Thumbnail Brief, and Youtube Thumbnail Tagline, using Number as the matching key.

8. Build thumbnail generation payload with Python

   Add a Code node (Python), read Thumbnail Brief and Tagline, and construct the JSON payload required by the Runware API.

   Execution prompt · Source text。Code node pythonCode field (Sticky Note3)

   Copy prompt

   import uuid
   
   thumbnail_brief = items[0]["json"].get("Youtube Thumbnail Brief", "")
   tagline = items[0]["json"].get("Youtube Thumbnail Tagline", "")
   
   prompt = (
       "Generate a professional YouTube thumbnail.\n\n"
       f"Thumbnail Context:\n{thumbnail_brief}\n\n"
       f"Main Tagline: '{tagline}'\n\n"
       "Thumbnail requirements:\n"
       "- Bold, central tagline text\n"
       "- Cinematic lighting, vibrant colors\n"
       "- Highly clickable YouTube style\n"
       "- 1280x720 resolution"
   )
   
   payload = {
       "taskType": "imageInference",
       "taskUUID": str(uuid.uuid4()),
       "positivePrompt": prompt,
       "width": 1280,
       "height": 720,
       "model": "bytedance:3@1",
       "numberResults": 1,
       "outputFileName": "thumbnail.png"
   }
   
   return {"data": [payload]}

9. Call Runware API to generate thumbnail

   Add an HTTP Request node, POST to https://api.runware.ai/v1/image, include Authorization and Content-Type: application/json headers, and pass the payload from the previous step as the body.

10. Write thumbnail URL back to Google Sheets

    Use a Google Sheets node to update the same row, writing the returned imageURL into the Youtube Generated Thumbnail column, using Number as the matching key.

You are an n8n workflow setup assistant. The user wants to reproduce a workflow that auto-generates YouTube SEO content and thumbnails. Assist in this order: 1) Confirm the user has an n8n instance, Google Sheets OAuth2 credentials, OpenAI API key, and Runware API key; 2) Guide them to create a Google Sheet with columns: Number, Roman Hindi Script, English script, Youtube Title, Youtube Tags, Youtube Keywords, Youtube Description, Youtube Thumbnail Brief, Youtube Thumbnail Tagline, Youtube Generated Thumbnail; 3) Walk through importing or manually configuring nodes; 4) Remind them to replace credential IDs and spreadsheet IDs; 5) Suggest testing with a single row first.