Génération de variantes d'images publicitaires avec GPT-4, Dumpling AI et Google Drive

• •Informations d'identification Google Drive API
• •Peut nécessiter les informations d'identification d'authentification de l'API cible
• •Informations d'identification Google Sheets API
• •Clé API OpenAI

{
  "id": "cU2fwVR3z955Vuqt",
  "meta": {
    "instanceId": "a1ae5c8dc6c65e674f9c3947d083abcc749ef2546dff9f4ff01de4d6a36ebfe6",
    "templateCredsSetupCompleted": true
  },
  "name": "Generate Ad Image Variations Using GPT-4, Dumpling AI & Google Drive",
  "tags": [],
  "nodes": [
    {
      "id": "849a3377-d9bb-42d4-b602-61ca789bfb1f",
      "name": "Soumettre les informations de la marque + image",
      "type": "n8n-nodes-base.formTrigger",
      "position": [
        240,
        55
      ],
      "webhookId": "da177997-c71a-4127-816e-d0e3a442b94a",
      "parameters": {
        "options": {},
        "formTitle": "Ad Image Generator",
        "formFields": {
          "values": [
            {
              "fieldLabel": "Brand Name",
              "requiredField": true
            },
            {
              "fieldLabel": "Brand Website",
              "requiredField": true
            },
            {
              "fieldType": "file",
              "fieldLabel": "Ad Image",
              "multipleFiles": false,
              "requiredField": true
            }
          ]
        },
        "formDescription": "Provide brand and product info, and an optional reference image to generate image variations with AI."
      },
      "typeVersion": 2.2
    },
    {
      "id": "25dc9e03-7c7e-44eb-897c-bd8956e554d7",
      "name": "Télécharger l'image publicitaire sur Drive Google",
      "type": "n8n-nodes-base.googleDrive",
      "position": [
        460,
        55
      ],
      "parameters": {
        "name": "={{ $json['Ad Image'].filename }} (Original)",
        "driveId": {
          "__rl": true,
          "mode": "list",
          "value": "My Drive",
          "cachedResultUrl": "https://drive.google.com/drive/my-drive",
          "cachedResultName": "My Drive"
        },
        "options": {},
        "folderId": {
          "__rl": true,
          "mode": "list",
          "value": "1R5bTxrKmi9NDMFJIh3aQgbNuZwmCybLV",
          "cachedResultUrl": "https://drive.google.com/drive/folders/1R5bTxrKmi9NDMFJIh3aQgbNuZwmCybLV",
          "cachedResultName": "n8n Testing"
        },
        "inputDataFieldName": "Ad_Image"
      },
      "credentials": {
        "googleDriveOAuth2Api": {
          "id": "B0Hulh0j065xznhI",
          "name": "nnekajennifer"
        }
      },
      "typeVersion": 3
    },
    {
      "id": "feb9a701-63d5-4c68-b4d2-56693dfd667a",
      "name": "Télécharger l'image publicitaire pour analyse",
      "type": "n8n-nodes-base.googleDrive",
      "position": [
        680,
        55
      ],
      "parameters": {
        "fileId": {
          "__rl": true,
          "mode": "id",
          "value": "={{ $(' Upload Ad Image to Google Drive').item.json.id }}"
        },
        "options": {},
        "operation": "download"
      },
      "credentials": {
        "googleDriveOAuth2Api": {
          "id": "B0Hulh0j065xznhI",
          "name": "nnekajennifer"
        }
      },
      "typeVersion": 3
    },
    {
      "id": "c6000bdb-ce69-4302-b27e-aab73d128337",
      "name": "Décrire le style visuel de l'image",
      "type": "@n8n/n8n-nodes-langchain.openAi",
      "position": [
        900,
        55
      ],
      "parameters": {
        "text": "Describe the visual style, subject matter, and composition of this image. Is it a lifestyle image, a product-only shot, or a combination? Include lighting style and camera angle if possible.",
        "modelId": {
          "__rl": true,
          "mode": "list",
          "value": "gpt-4o",
          "cachedResultName": "GPT-4O"
        },
        "options": {},
        "resource": "image",
        "inputType": "base64",
        "operation": "analyze"
      },
      "credentials": {
        "openAiApi": {
          "id": "dd8NvMC6rvx8RITo",
          "name": "OpenAi account 2"
        }
      },
      "typeVersion": 1.8
    },
    {
      "id": "110842b9-d0d0-4c46-bc2d-94851bd6aa30",
      "name": "Analyser le style du site web de la marque",
      "type": "@n8n/n8n-nodes-langchain.openAi",
      "position": [
        1120,
        55
      ],
      "parameters": {
        "modelId": {
          "__rl": true,
          "mode": "list",
          "value": "gpt-4",
          "cachedResultName": "GPT-4"
        },
        "options": {},
        "messages": {
          "values": [
            {
              "content": "=You are a visual brand strategist and art director for direct-to-consumer (DTC) e-commerce brands.\n\nPlease analyze the following brand website. Focus **only** on the brand’s **visual aesthetic**, including:\n\n- Color palette\n- Photography style and lighting\n- Imagery themes (e.g. lifestyle vs. product shots)\n- Mood or tone evoked by visuals\n- Any repeating design elements or layout patterns\n\nBe descriptive but concise. The output will be used to help design consistent and creative AI-generated images for ad creatives — so focus entirely on the visual look and feel.\n\n\nBrand Website: {{ $('Submit Brand Info + Image').item.json['Brand Website'] }}\nBrand Name: {{ $('Submit Brand Info + Image').item.json['Brand Name'] }}\n"
            }
          ]
        }
      },
      "credentials": {
        "openAiApi": {
          "id": "dd8NvMC6rvx8RITo",
          "name": "OpenAi account 2"
        }
      },
      "typeVersion": 1.8
    },
    {
      "id": "41a97d61-1d3a-4ec3-9b76-c5fe4fb2b312",
      "name": "Agent LangChain : Générer des prompts de variation",
      "type": "@n8n/n8n-nodes-langchain.agent",
      "position": [
        1496,
        55
      ],
      "parameters": {
        "text": "=Brand Name: {{ $('Submit Brand Info + Image').item.json['Brand Name'] }}\nWebsite: {{ $('Submit Brand Info + Image').item.json['Brand Website'] }}\nReference Ad Description: {{ $('Describe Visual Style of Image').item.json.content }}\nVisual Style Overview: {{ $json.message.content }}",
        "options": {
          "systemMessage": "=Brand Name: {{ $('Submit Brand Info + Image').item.json['Brand Name'] }}\nWebsite: {{ $('Submit Brand Info + Image').item.json['Brand Website'] }}\nReference Ad Description: {{ $('Describe Visual Style of Image').item.json.content }}\nVisual Style Overview: {{ $json.message.content }}\n\nYour task is to generate **10 tightly related visual variations** of a reference ad image — not new concepts. These prompts are for testing subtle creative changes on Facebook/Instagram ads (e.g., backdrop, mood, color, lighting) while preserving the original subject and composition.\n\n### Reference Materials:\n\n**Brand Info**: {{ $('Submit Brand Info + Image').item.json['Brand Name'] }}\n**Website**:  {{ $('Submit Brand Info + Image').item.json['Brand Website'] }}\n**Visual Aesthetics Guideline**: {{ $json.message.content }}\n**Reference Image Description**: {{ $('Describe Visual Style of Image').item.json.content }}\n\nUse the Reference Image Description to ground your concepts in visual reality — the new prompts should **feel like believable variations** of this original ad image in terms of camera angle, lighting, and context, while introducing fresh concepts or creative twists.\n\n### For each of the 10 prompts:\n\n- **Preserve** the reference image's subject (product, camera angle, core framing)\n- **Only vary** the background, environment, mood, lighting, or color treatment\n- Examples of variation types:\n  - Daylight vs sunset lighting\n  - Poolside vs marble countertop\n  - Lavender tones vs beach sand tones\n  - Summer vibe vs spa-like calm\n- Use vivid, sensory language to describe each variation\n- Always include **aspect ratio** (1:1 or 4:5)\n- Never introduce logos, overlays, or major subject changes\n\nThe goal is to create **subtle, performance-testable image variations**, not entirely new compositions.\n\nPlease return **only** a JSON array of 10 objects, each with a single property `\"prompt\"` containing the image prompt. Example output structure:\n\n[\n  {\n    \"prompt\": \"Sun-drenched poolside shot of the product on a marble ledge at golden hour, with soft shadows and warm tones. Aspect ratio 1:1.\"\n  },\n  {\n    \"prompt\": \"Cool lavender-tinted sunset beach backdrop behind the product, highlighting reflective metallic accents. Aspect ratio 4:5.\"\n  },\n  {\n    \"prompt\": \"...\"\n  }\n]"
        },
        "promptType": "define",
        "hasOutputParser": true
      },
      "typeVersion": 1.7
    },
    {
      "id": "549abb3a-c944-47ce-a5eb-a5e35e8b7934",
      "name": "Parser les prompts en tableau JSON",
      "type": "@n8n/n8n-nodes-langchain.outputParserStructured",
      "position": [
        1720,
        260
      ],
      "parameters": {
        "jsonSchemaExample": "[\n  {\n    \"prompt\": \"Sun-drenched poolside shot of the product on a marble ledge at golden hour, with soft shadows and warm tones. Aspect ratio 1:1.\"\n  },\n  {\n    \"prompt\": \"Cool lavender-tinted sunset beach backdrop behind the product, highlighting reflective metallic accents. Aspect ratio 4:5.\"\n  },\n  {\n    \"prompt\": \"...\"\n  }\n]"
      },
      "typeVersion": 1.2
    },
    {
      "id": "833fd095-e1ea-4b70-989f-542b415f4294",
      "name": "Diviser : Un prompt par élément",
      "type": "n8n-nodes-base.splitOut",
      "position": [
        1872,
        55
      ],
      "parameters": {
        "options": {},
        "fieldToSplitOut": "output"
      },
      "typeVersion": 1
    },
    {
      "id": "89875b01-68e0-4aec-a688-9965a8ca4a3d",
      "name": "Télécharger l'image de base pour chaque variation",
      "type": "n8n-nodes-base.googleDrive",
      "position": [
        2092,
        55
      ],
      "parameters": {
        "fileId": {
          "__rl": true,
          "mode": "id",
          "value": "={{ $(' Upload Ad Image to Google Drive').item.json.id }}"
        },
        "options": {},
        "operation": "download"
      },
      "credentials": {
        "googleDriveOAuth2Api": {
          "id": "B0Hulh0j065xznhI",
          "name": "nnekajennifer"
        }
      },
      "typeVersion": 3
    },
    {
      "id": "6c0131d3-46c3-47af-b8d4-d293f5378760",
      "name": "Boucle : Traiter les variations d'images",
      "type": "n8n-nodes-base.splitInBatches",
      "position": [
        2312,
        55
      ],
      "parameters": {
        "options": {}
      },
      "typeVersion": 3
    },
    {
      "id": "046f2619-c1a3-4674-9ae6-33221bb6f095",
      "name": "Dumpling AI : Générer une variation d'image",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        2532,
        -20
      ],
      "parameters": {
        "url": "https://app.dumplingai.com/api/v1/generate-ai-image",
        "method": "POST",
        "options": {},
        "jsonBody": "={\n  \"model\": \"FLUX.1-pro\",\n  \"input\": {\n    \"prompt\": \"{{ $json.prompt }}\"\n  }\n}\n",
        "sendBody": true,
        "specifyBody": "json",
        "authentication": "genericCredentialType",
        "genericAuthType": "httpHeaderAuth"
      },
      "credentials": {
        "httpHeaderAuth": {
          "id": "6X2b4RLVRbz8I4Z8",
          "name": "Vapi"
        }
      },
      "typeVersion": 4.2
    },
    {
      "id": "815e004d-f1ed-485f-be24-c4bf565a61f1",
      "name": "Enregistrer les URLs des variations d'images dans Sheets Google",
      "type": "n8n-nodes-base.googleSheets",
      "position": [
        2752,
        55
      ],
      "parameters": {
        "columns": {
          "value": {
            "Image URL": "={{ $json.url }}"
          },
          "schema": [
            {
              "id": "Image URL",
              "type": "string",
              "display": true,
              "removed": false,
              "required": false,
              "displayName": "Image URL",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            }
          ],
          "mappingMode": "defineBelow",
          "matchingColumns": [
            "Image URL"
          ],
          "attemptToConvertTypes": false,
          "convertFieldsToString": false
        },
        "options": {},
        "operation": "append",
        "sheetName": {
          "__rl": true,
          "mode": "list",
          "value": "gid=0",
          "cachedResultUrl": "https://docs.google.com/spreadsheets/d/12bBSaMdX-jE7QtIlEX4CBUEGScmJx3xK7bwFtHWan64/edit#gid=0",
          "cachedResultName": "Sheet1"
        },
        "documentId": {
          "__rl": true,
          "mode": "list",
          "value": "",
          "cachedResultUrl": "https://docs.google.com/spreadsheets/d/12bBSaMdX-jE7QtIlEX4CBUEGScmJx3xK7bwFtHWan64/edit?usp=drivesdk",
          "cachedResultName": "Edited image "
        }
      },
      "credentials": {
        "googleSheetsOAuth2Api": {
          "id": "GaJqJHuS5mQxap7q",
          "name": "Google Sheets account"
        }
      },
      "typeVersion": 4.6
    },
    {
      "id": "165b9120-6c0b-446e-afd6-c4f233d167df",
      "name": "GPT-4o (Connecté à l'agent LangChain)",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
      "position": [
        1524,
        275
      ],
      "parameters": {
        "model": {
          "__rl": true,
          "mode": "list",
          "value": "gpt-4o-mini"
        },
        "options": {}
      },
      "credentials": {
        "openAiApi": {
          "id": "dd8NvMC6rvx8RITo",
          "name": "OpenAi account 2"
        }
      },
      "typeVersion": 1.2
    },
    {
      "id": "b52a1b6b-d5c5-43ad-a2c4-b318aa715f7d",
      "name": "Note autocollante",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        300,
        -480
      ],
      "parameters": {
        "width": 780,
        "height": 740,
        "content": "### 🖼️ Ad Image Variation Generator (Using GPT-4 + Dumpling AI)\n\nThis workflow creates 10 subtle creative variations of a reference ad image  \nto test performance across visual styles, lighting, background, and tone —  \nwhile preserving the product's framing and subject.\n\n---\n\n### 🔧 How It Works\n\n1. User submits brand name, website, and reference ad image via a form.\n2. The image is uploaded to Google Drive.\n3. GPT-4o analyzes the image’s visual style (composition, subject, lighting).\n4. GPT-4 analyzes the brand website to understand overall visual identity.\n5. A LangChain AI Agent uses both to generate 10 variation prompts.\n6. Each prompt is passed to Dumpling AI to generate a new ad image.\n7. All image URLs are logged in Google Sheets.\n\n"
      },
      "typeVersion": 1
    }
  ],
  "active": false,
  "pinData": {},
  "settings": {
    "executionOrder": "v1"
  },
  "versionId": "3bc41ee0-40f8-4a07-b76d-e8fe2ff5a0c8",
  "connections": {
    "849a3377-d9bb-42d4-b602-61ca789bfb1f": {
      "main": [
        [
          {
            "node": "25dc9e03-7c7e-44eb-897c-bd8956e554d7",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "833fd095-e1ea-4b70-989f-542b415f4294": {
      "main": [
        [
          {
            "node": "89875b01-68e0-4aec-a688-9965a8ca4a3d",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "110842b9-d0d0-4c46-bc2d-94851bd6aa30": {
      "main": [
        [
          {
            "node": "41a97d61-1d3a-4ec3-9b76-c5fe4fb2b312",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "549abb3a-c944-47ce-a5eb-a5e35e8b7934": {
      "ai_outputParser": [
        [
          {
            "node": "41a97d61-1d3a-4ec3-9b76-c5fe4fb2b312",
            "type": "ai_outputParser",
            "index": 0
          }
        ]
      ]
    },
    "c6000bdb-ce69-4302-b27e-aab73d128337": {
      "main": [
        [
          {
            "node": "110842b9-d0d0-4c46-bc2d-94851bd6aa30",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "feb9a701-63d5-4c68-b4d2-56693dfd667a": {
      "main": [
        [
          {
            "node": "c6000bdb-ce69-4302-b27e-aab73d128337",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "6c0131d3-46c3-47af-b8d4-d293f5378760": {
      "main": [
        [],
        [
          {
            "node": "046f2619-c1a3-4674-9ae6-33221bb6f095",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "25dc9e03-7c7e-44eb-897c-bd8956e554d7": {
      "main": [
        [
          {
            "node": "feb9a701-63d5-4c68-b4d2-56693dfd667a",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "046f2619-c1a3-4674-9ae6-33221bb6f095": {
      "main": [
        [
          {
            "node": "815e004d-f1ed-485f-be24-c4bf565a61f1",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "165b9120-6c0b-446e-afd6-c4f233d167df": {
      "ai_languageModel": [
        [
          {
            "node": "41a97d61-1d3a-4ec3-9b76-c5fe4fb2b312",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "89875b01-68e0-4aec-a688-9965a8ca4a3d": {
      "main": [
        [
          {
            "node": "6c0131d3-46c3-47af-b8d4-d293f5378760",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "815e004d-f1ed-485f-be24-c4bf565a61f1": {
      "main": [
        [
          {
            "node": "6c0131d3-46c3-47af-b8d4-d293f5378760",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "41a97d61-1d3a-4ec3-9b76-c5fe4fb2b312": {
      "main": [
        [
          {
            "node": "833fd095-e1ea-4b70-989f-542b415f4294",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}