Telegram을 사용하여 LLM을 강화한 트리거와 Gemini/Leonardo.AI를 사용하여 애니메이션 이미지 생성

• •Telegram Bot Token
• •대상 API의 인증 정보가 필요할 수 있음

{
  "meta": {
    "instanceId": "702596a9719c2ee1ff81772a5069e7af986ac325ce905b7dfad8a8c8c8c9c76c"
  },
  "nodes": [
    {
      "id": "0290ceba-fc2d-426b-bcb3-4caa37b4d690",
      "name": "메모",
      "type": "n8n-nodes-base.stickyNote",
      "disabled": true,
      "position": [
        0,
        -240
      ],
      "parameters": {
        "color": 5,
        "width": 1220,
        "height": 92,
        "content": "# Image generation with Gemini [Free for 90-days with a GCP acc]"
      },
      "typeVersion": 1
    },
    {
      "id": "f9b5e427-a030-4323-b8fb-69928e2b44ae",
      "name": "Telegram 트리거",
      "type": "n8n-nodes-base.telegramTrigger",
      "position": [
        64,
        192
      ],
      "webhookId": "cdfeedc2-f811-4dbe-bca7-863ac47dd841",
      "parameters": {
        "updates": [
          "message"
        ],
        "additionalFields": {
          "chatIds": ""
        }
      },
      "credentials": {
        "telegramApi": {
          "id": "DCT0rTb6lDEXngOg",
          "name": "LXirein Bot (SMS)"
        }
      },
      "typeVersion": 1.2
    },
    {
      "id": "f2edb1fd-dfd1-48d5-a20d-5c48a603c9b0",
      "name": "HTTP- Gemini",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        1424,
        320
      ],
      "parameters": {
        "url": "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash-preview-image-generation:generateContent",
        "method": "POST",
        "options": {},
        "jsonBody": "={\n  \"contents\": [\n    {\n      \"parts\": [\n        {\n          \"text\": {{ JSON.stringify($json.description) }}\n        }\n      ]\n    }\n  ],\n  \"generationConfig\": {\n    \"responseModalities\": [\n      \"TEXT\",\n      \"IMAGE\"\n    ]\n  }\n}",
        "sendBody": true,
        "specifyBody": "json",
        "authentication": "predefinedCredentialType",
        "nodeCredentialType": "googlePalmApi"
      },
      "credentials": {
        "googlePalmApi": {
          "id": "o1FK0fi6w4RpYPp2",
          "name": "Gemini(PaLM) [api.Gmail]"
        }
      },
      "typeVersion": 4.2
    },
    {
      "id": "f117d6fd-bcbd-4ce9-b10c-3810e603eacd",
      "name": "설정 base64",
      "type": "n8n-nodes-base.set",
      "position": [
        1648,
        320
      ],
      "parameters": {
        "options": {},
        "assignments": {
          "assignments": [
            {
              "id": "31a2b3db-78a0-41b8-815f-ffb6c572279d",
              "name": "base64",
              "type": "string",
              "value": "={{\n  $json.candidates && $json.candidates[0] && $json.candidates[0].content && $json.candidates[0].content.parts && (\n    ($json.candidates[0].content.parts[1] && $json.candidates[0].content.parts[1].inlineData && $json.candidates[0].content.parts[1].inlineData.data) ||\n    ($json.candidates[0].content.parts[0] && $json.candidates[0].content.parts[0].inlineData && $json.candidates[0].content.parts[0].inlineData.data)\n  )\n}}"
            }
          ]
        }
      },
      "typeVersion": 3.4
    },
    {
      "id": "487b7016-35d2-459f-b1d3-5e16fcaa5681",
      "name": "대기 10s",
      "type": "n8n-nodes-base.wait",
      "position": [
        2080,
        320
      ],
      "webhookId": "74276acc-e402-42a8-9cd5-dffc3cb51c58",
      "parameters": {
        "amount": 10
      },
      "typeVersion": 1.1
    },
    {
      "id": "226d63d1-6e09-46e8-abd4-ad6d9b63a939",
      "name": "Image-count",
      "type": "n8n-nodes-base.set",
      "position": [
        304,
        192
      ],
      "parameters": {
        "options": {},
        "assignments": {
          "assignments": [
            {
              "id": "6c546bd1-389e-47e3-bd85-58a37750b6a2",
              "name": "User_Prompt",
              "type": "string",
              "value": "={{ $json.message.text }}"
            },
            {
              "id": "d15f7ae7-5203-4f48-9d4a-357444253c89",
              "name": "num_of_Images",
              "type": "number",
              "value": 4
            }
          ]
        }
      },
      "typeVersion": 3.4
    },
    {
      "id": "376700cb-8da4-42b3-b5db-527a336040bd",
      "name": "Loop Over Prompts",
      "type": "n8n-nodes-base.splitInBatches",
      "position": [
        1104,
        192
      ],
      "parameters": {
        "options": {}
      },
      "typeVersion": 3
    },
    {
      "id": "41a30836-c00f-49ae-ade1-7ad04a4a7a96",
      "name": "파일로 변환",
      "type": "n8n-nodes-base.convertToFile",
      "position": [
        1872,
        320
      ],
      "parameters": {
        "options": {},
        "operation": "toBinary",
        "sourceProperty": "base64"
      },
      "typeVersion": 1.1
    },
    {
      "id": "59590e2f-7150-4c12-988a-577b81f9c515",
      "name": "메모1",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1344,
        224
      ],
      "parameters": {
        "color": 4,
        "width": 260,
        "height": 300,
        "content": "## Using gemini model\n## to generate image"
      },
      "typeVersion": 1
    },
    {
      "id": "117c00f6-14fd-4a78-ae8c-f71e15a6016b",
      "name": "HTTP- Leonardo AI",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        1360,
        704
      ],
      "parameters": {
        "url": "https://cloud.leonardo.ai/api/rest/v1/generations",
        "method": "POST",
        "options": {},
        "sendBody": true,
        "authentication": "genericCredentialType",
        "bodyParameters": {
          "parameters": [
            {
              "name": "height",
              "value": "768"
            },
            {
              "name": "modelId",
              "value": "e71a1c2f-4f80-4800-934f-2c68979d8cc8"
            },
            {
              "name": "num_images",
              "value": "1"
            },
            {
              "name": "prompt",
              "value": "={{ JSON.stringify($json.description) }}"
            },
            {
              "name": "width",
              "value": "1024"
            },
            {
              "name": "public",
              "value": "true"
            },
            {
              "name": "promptMagic",
              "value": "true"
            },
            {
              "name": "guidance_scale",
              "value": "7"
            },
            {
              "name": "alchemy",
              "value": "true"
            },
            {
              "name": "photoReal",
              "value": "false"
            },
            {
              "name": "presetStyle",
              "value": "ANIME"
            }
          ]
        },
        "genericAuthType": "httpHeaderAuth"
      },
      "credentials": {
        "httpHeaderAuth": {
          "id": "b8Ao8Jb2KoJezTx5",
          "name": "Relevance.ai [cocaina.Gmail]"
        }
      },
      "typeVersion": 4.2
    },
    {
      "id": "cbc60c2f-5f6e-4ff7-8afd-69559c9746d6",
      "name": "분할 출력 Prompts",
      "type": "n8n-nodes-base.splitOut",
      "position": [
        880,
        192
      ],
      "parameters": {
        "options": {},
        "fieldToSplitOut": "output"
      },
      "typeVersion": 1
    },
    {
      "id": "7b9caa6d-84e2-4014-8fac-ff8b9bb84a6f",
      "name": "메모2",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1264,
        592
      ],
      "parameters": {
        "color": 4,
        "width": 440,
        "height": 300,
        "content": "## Using Leonardo.AI to generate image\nReplace `HTTP- Gemini` node with this node.\nCreate new header auth credential with your leonardo.ai API key."
      },
      "typeVersion": 1
    },
    {
      "id": "25e7606d-3adc-466a-861e-9b2640be6578",
      "name": "구조화된 출력 파서",
      "type": "@n8n/n8n-nodes-langchain.outputParserStructured",
      "position": [
        704,
        400
      ],
      "parameters": {
        "schemaType": "manual",
        "inputSchema": "{\n  \"type\": \"array\",\n  \"items\": {\n    \"type\": \"object\",\n    \"properties\": {\n      \"description\": {\n        \"type\": \"string\",\n        \"description\": \"The full, detailed prompt for the anime background, including artistic style, lighting, mood, and specific elements, suitable for an AI image model.\"\n      }\n    },\n    \"required\": [\"description\"]\n  }\n}"
      },
      "typeVersion": 1.2
    },
    {
      "id": "0dd62df2-6ffe-4d20-a327-d0aaa383aad1",
      "name": "Prompt generator (USER_Inspired)",
      "type": "@n8n/n8n-nodes-langchain.chainLlm",
      "position": [
        512,
        192
      ],
      "parameters": {
        "text": "=Enhance My Anime Image Prompt!\n\nI have a basic idea for an anime image, but I need your help to make it more detailed and visually appealing for AI image generation.\nPlease take my simple prompt below and transform it into *{{ $('Image-count').item.json.num_of_Images }}* highly detailed and compelling prompts. Make sure to include a rich character description, a vivid background description, specific artistic style instructions, and an emotional/narrative hook. Keep the output moderately detailed, focusing on impactful elements.\n\nHere's my amateur prompt:\n{{ $('Image-count').item.json.User_Prompt }}",
        "messages": {
          "messageValues": [
            {
              "message": "=You are an expert anime artist and prompt generator for AI image generation models. Your goal is to take a user's initial, amateur prompt for an anime character or scene and transform it into a highly detailed, compelling, and visually attractive prompt. The refined prompt should resonate deeply with anime lovers and the general public, focusing on visual appeal, emotional impact, and storytelling potential.\n\n\n***Input: You will receive a user-provided prompt, which may be brief or lack specific detail.***\n\n***Output: Generate one highly refined and enhanced anime character and background prompt. This prompt must include:***\n\n* A highly detailed character description: Elaborating on appearance, clothing, accessories, and a clear expression of personality/mood.\n* A highly detailed background description: Richly setting the scene, time of day, atmosphere, lighting, and key elements, ensuring it complements the character.\n* Specific artistic style and rendering instructions: To ensure a polished, professional anime look, including terms like \"masterpiece,\" \"trending on ArtStation,\" and specific anime studio or artist influences if appropriate.\n* Emotional resonance/narrative hook: Something that draws the viewer in and suggests a deeper story or feeling.\n\n\n\n***Chain of Thought for Refining Prompts:***\n\n\n**Deconstruct User Input:**\n\n* Identify the core character concept (e.g., \"girl,\" \"hero,\" \"student\").\n* Identify the core background concept (e.g., \"city,\" \"nature,\" \"room\").\n* Note any existing details, however sparse (e.g., \"blue hair,\" \"sunset\").\n* Assess the implied mood or genre (e.g., \"cute,\" \"sad,\" \"action\").\n\n**Expand Character Details:**\n\n* Appearance: Add specific hair color, style, length (e.g., \"long, flowing auburn hair,\" \"short, spiky black hair with undercut\"). Define eye color and shape (e.g., \"large, expressive emerald eyes,\" \"sharp, piercing golden eyes\").\n* Clothing: Go beyond generic descriptions. Specify fabric, style, layering, accessories (e.g., \"oversized, cable-knit cardigan in cream,\" \"sleek, futuristic battle armor with glowing energy conduits,\" \"vintage floral sundress with lace trim\").\n* Personality/Mood: Infuse subtle cues through expression, pose, or implied action (e.g., \"gentle, contemplative gaze,\" \"playful smirk,\" \"determined stance\"). Consider what they might be doing or thinking.\n\n** Elaborate Background Details:**\n\n* Scene: Expand on the type of location (e.g., \"bustling Shibuya crossing at night,\" \"serene bamboo grove with a hidden shrine,\" \"cozy, book-filled attic room with a skylight\").\n* Time/Atmosphere: Specify time of day and how it impacts lighting (e.g., \"golden hour with long, soft shadows,\" \"moonlit night with shimmering reflections,\" \"overcast morning with a cool, muted palette\").\n* Key Elements: Add specific objects, weather, or environmental effects (e.g., \"gentle rain leaving reflections on wet pavement,\" \"steam rising from manholes,\" \"cherry blossoms drifting through the air,\" \"ancient stone lanterns glowing softly\").\n* Depth & Perspective: Consider foreground, midground, and background elements.\n\n\n** Inject Artistic Style & Rendering Instructions:**\n\n* Translate the implied genre or mood into specific anime art styles (e.g., \"Makoto Shinkai-esque for atmospheric beauty,\" \"Kyoto Animation style for expressive characters,\" \"dynamic shonen art for action\").\n* Add quality descriptors: \"masterpiece,\" \"ultra-detailed,\" \"cinematic lighting,\" \"volumetric lighting,\" \"trending on ArtStation,\" \"concept art,\" \"8K.\"\n* Suggest color palettes: \"vibrant neon colors,\" \"soft pastel hues,\" \"muted earth tones.\"\n\n\n**Enhance Emotional/Narrative Hook:**\n\n* Consider the feeling the image should evoke (e.g., \"nostalgia,\" \"wonder,\" \"solitude,\" \"excitement,\" \"peace\").\n* Suggest a subtle narrative: Is the character waiting? Reflecting? Embarking on a journey? Interacting with the environment?\n\n\n***Review and Optimize:***\n\n* Read the complete prompt aloud. Does it flow well?\n* Is it specific enough for an AI to interpret accurately?\n* Is it evocative and appealing?\n* Are there strong adjectives and verbs?\n* Remove redundancies.\n* Ensure consistency between character and background mood.\n\n\n**Example Prompts (Your output should follow this structure):**\n\n*Example 1: Schoolgirl at Sunset*\nUser's Amateur Prompt:\n\"Anime girl in a school uniform at sunset.\"\n\nLLM's Moderately Detailed Prompt:\nA vibrant anime illustration of a cheerful young girl with shoulder-length brown hair and bright, sparkling blue eyes, wearing a standard Japanese school uniform (white blouse, navy blazer, red ribbon). She stands with a gentle smile, holding her school bag.\nThe background is a serene school courtyard at golden hour sunset. Warm, soft light bathes the brick building and illuminates the rustling leaves of a nearby cherry tree. Long shadows stretch across the ground, and the sky displays a beautiful gradient of orange, pink, and soft purple hues. A sense of peaceful nostalgia fills the air.\nArt Style: Kyoto Animation style, warm color palette, detailed lighting, masterpiece, trending on ArtStation.\n\n\n*Example 2: Fantasy Adventurer*\nUser's Amateur Prompt:\n\"Fantasy guy with a sword in a cave.\"\n\nLLM's Moderately Detailed Prompt:\nA dynamic anime illustration of a determined young man with short, spiky black hair and intense, glowing green eyes. He wears worn leather armor over a dark tunic and wields a glowing runic sword in a ready stance. His expression is focused and resolute.\nThe background is a dark, mystical cave illuminated by bioluminescent crystals. Faint blue and green light emanates from glowing formations on the damp rock walls. A small, cascading waterfall creates a soft mist in the air. The atmosphere is mysterious and adventurous, with hints of ancient magic.\nArt Style: Dynamic action anime, dark fantasy elements, volumetric lighting, highly detailed, concept art, trending on ArtStation.\n\n\n*Example 3: Cozy Café Scene*\nUser's Amateur Prompt:\n\"Anime character in a cafe.\"\n\nLLM's Moderately Detailed Prompt:\nA charming anime illustration of a relaxed young woman with long, wavy blonde hair and warm, inviting brown eyes. She sips from a steaming coffee mug, wearing a comfortable, oversized beige sweater. Her expression is one of calm contentment.\nThe background is a cozy, warmly lit café interior on a rainy afternoon. Soft light filters through the window, highlighting gentle raindrops sliding down the glass. Bookshelves filled with eclectic titles line the walls, and a few potted plants add warmth. The ambiance is peaceful and inviting, perfect for quiet reflection.\nArt Style: Slice-of-life anime, soft lighting, warm color palette, aesthetically pleasing, masterpiece, comforting mood."
            }
          ]
        },
        "promptType": "define",
        "hasOutputParser": true
      },
      "typeVersion": 1.5
    },
    {
      "id": "41523fb6-2de7-4123-9276-f12e906f94d1",
      "name": "DeepSeek from OpenRouter",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenRouter",
      "position": [
        480,
        400
      ],
      "parameters": {
        "model": "deepseek/deepseek-chat-v3-0324:free",
        "options": {
          "topP": 0.8,
          "temperature": 0.7
        }
      },
      "credentials": {
        "openRouterApi": {
          "id": "oPGlise3ZkYmOynS",
          "name": "OpenRouter [Api.Gmail]"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "bf7ec591-bba2-4f11-9139-311e5040acc6",
      "name": "메모4",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -640,
        -224
      ],
      "parameters": {
        "color": 6,
        "width": 598,
        "height": 1280,
        "content": "# 🎨 Anime Image Generator — Setup & Notes\n\nThis workflow turns a **simple text prompt** into **high-quality anime-style images** using AI.\n\n---\n\n## ⚙️ How It Works\n1. **Telegram Trigger** → user sends a text prompt.  \n2. **Prompt Generator (LLM)** → refines the amateur prompt into multiple rich, vivid anime prompts.  \n3. **Image Generation** → \n   - Default: **Gemini** (90-day free trial with GCP account).  \n   - Optional: **Leonardo.AI API** (better quality, paid).  \n4. **Conversion & Delivery** → converts base64 to file and sends final images back via Telegram.\n\n---\n\n## 🔑 Setup Steps\n- **Telegram Bot:**  \n  - Create a bot with [BotFather](https://t.me/BotFather).  \n  - Add your Bot Token and your Telegram User ID to the workflow.\n\n- **Choose Image Model:**  \n  - Start with **Gemini** for free trial.  \n  - Switch to **Leonardo.AI** for higher-quality anime results (replace the `HTTP-Gemini` node with `HTTP-Leonardo AI` and add your Leonardo API key as a Header Auth credential).\n\n- **LLM Provider:**  \n  - Uses **DeepSeek via OpenRouter** (can swap with other chat LLMs).  \n  - Handles prompt enhancement for vivid anime-style scenes.\n\n- **Optional Adjustments:**  \n  - Change the default **number of images** in the `Image-count` node.  \n  - Modify artistic style (e.g., pastel slice-of-life, fantasy action) by tweaking the prompt template.\n\n---\n\n## 💡 Tips\n- Start with **short, simple prompts** — e.g., *“girl with umbrella in neon city street at night”*.  \n  The LLM expands them into rich anime prompts.\n- If output looks flat, experiment with **Leonardo.AI** and adjust guidance / style parameters.\n- Avoid NSFW or copyrighted character requests — follow provider policies.\n- Keep API keys **private** and never share them in screenshots.\n\n---\n\n✨ You’re ready to create **stunning anime images** directly from Telegram!\n"
      },
      "typeVersion": 1
    },
    {
      "id": "b060f972-79f4-42dd-8def-2acc3b63eb47",
      "name": "메모5",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        2016,
        192
      ],
      "parameters": {
        "width": 256,
        "height": 288,
        "content": "## Not necessary\nUsed to avoid rate limits"
      },
      "typeVersion": 1
    },
    {
      "id": "1c590870-8030-4b6f-85a8-7ccf5c78c8e9",
      "name": "메모7",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        16,
        64
      ],
      "parameters": {
        "color": 3,
        "height": 256,
        "content": "## Configure Me!\nPaste your own chatID in the \"Restrict to Chat IDs\" field"
      },
      "typeVersion": 1
    },
    {
      "id": "6b04e902-842e-4550-b6b0-e94decd4a859",
      "name": "Send pics via Telegram",
      "type": "n8n-nodes-base.telegram",
      "position": [
        1504,
        -48
      ],
      "webhookId": "fb21d8df-f012-4bdb-908f-457610cbe930",
      "parameters": {
        "chatId": "=",
        "operation": "sendPhoto",
        "binaryData": true,
        "additionalFields": {}
      },
      "credentials": {
        "telegramApi": {
          "id": "DCT0rTb6lDEXngOg",
          "name": "LXirein Bot (SMS)"
        }
      },
      "typeVersion": 1.2
    },
    {
      "id": "72ba948d-eb4c-4ee1-ba22-a441ddfcec4d",
      "name": "메모8",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1440,
        -160
      ],
      "parameters": {
        "color": 3,
        "height": 256,
        "content": "## Configure Me!\nPaste your own chatID in the \"ChatID\" field"
      },
      "typeVersion": 1
    },
    {
      "id": "e8584025-bf4d-44af-8680-00b28a39bda0",
      "name": "메모6",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        528,
        128
      ],
      "parameters": {
        "width": 704,
        "height": 80,
        "content": "## Generates 4 images for an image query"
      },
      "typeVersion": 1
    },
    {
      "id": "f417031a-195f-409c-9584-02f4e9844f11",
      "name": "메모3",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1200,
        864
      ],
      "parameters": {
        "width": 544,
        "height": 80,
        "content": "You cn try using different models for image generation here.\nGemini - [Good quality, 90 free trial with GCP account] (With nano-banana)\nLeonardo.AI - [Better quality, Paid API]\n"
      },
      "typeVersion": 1
    }
  ],
  "pinData": {},
  "connections": {
    "Wait 10s": {
      "main": [
        [
          {
            "node": "376700cb-8da4-42b3-b5db-527a336040bd",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Set base64": {
      "main": [
        [
          {
            "node": "Convert to File",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "226d63d1-6e09-46e8-abd4-ad6d9b63a939": {
      "main": [
        [
          {
            "node": "0dd62df2-6ffe-4d20-a327-d0aaa383aad1",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "f2edb1fd-dfd1-48d5-a20d-5c48a603c9b0": {
      "main": [
        [
          {
            "node": "Set base64",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Convert to File": {
      "main": [
        [
          {
            "node": "Wait 10s",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Telegram Trigger": {
      "main": [
        [
          {
            "node": "226d63d1-6e09-46e8-abd4-ad6d9b63a939",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "376700cb-8da4-42b3-b5db-527a336040bd": {
      "main": [
        [
          {
            "node": "6b04e902-842e-4550-b6b0-e94decd4a859",
            "type": "main",
            "index": 0
          }
        ],
        [
          {
            "node": "f2edb1fd-dfd1-48d5-a20d-5c48a603c9b0",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Split Out Prompts": {
      "main": [
        [
          {
            "node": "376700cb-8da4-42b3-b5db-527a336040bd",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "41523fb6-2de7-4123-9276-f12e906f94d1": {
      "ai_languageModel": [
        [
          {
            "node": "0dd62df2-6ffe-4d20-a327-d0aaa383aad1",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "Structured Output Parser": {
      "ai_outputParser": [
        [
          {
            "node": "0dd62df2-6ffe-4d20-a327-d0aaa383aad1",
            "type": "ai_outputParser",
            "index": 0
          }
        ]
      ]
    },
    "0dd62df2-6ffe-4d20-a327-d0aaa383aad1": {
      "main": [
        [
          {
            "node": "Split Out Prompts",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}