{
  "name": "AI Social Media Content Generator (Ollama)",
  "nodes": [
    {
      "parameters": {},
      "id": "trigger",
      "name": "Manual Trigger",
      "type": "n8n-nodes-base.manualTrigger",
      "typeVersion": 1,
      "position": [
        240,
        300
      ]
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "topic",
              "name": "topic",
              "value": "Benefits of self-hosting AI models",
              "type": "string"
            },
            {
              "id": "brand_voice",
              "name": "brand_voice",
              "value": "Tech-savvy, slightly irreverent, helpful. We love open source and privacy.",
              "type": "string"
            },
            {
              "id": "target_audience",
              "name": "target_audience",
              "value": "developers, sysadmins, tech enthusiasts",
              "type": "string"
            }
          ]
        }
      },
      "id": "config",
      "name": "Content Parameters",
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        460,
        300
      ]
    },
    {
      "parameters": {
        "promptType": "define",
        "text": "=Generate social media content for the topic: \"{{ $json.topic }}\"\n\nBrand voice: {{ $json.brand_voice }}\nTarget audience: {{ $json.target_audience }}\n\nReturn ONLY valid JSON with this structure:\n{\n  \"twitter\": \"Tweet (max 280 chars, include 2-3 relevant hashtags)\",\n  \"linkedin\": \"LinkedIn post (3-4 paragraphs, professional tone, include hashtags)\",\n  \"reddit_title\": \"Reddit post title (engaging, not clickbait)\",\n  \"reddit_body\": \"Reddit post body (informative, genuine value, no hard sell)\",\n  \"instagram_caption\": \"Instagram caption (engaging, with emoji, 5-10 hashtags at end)\"\n}",
        "batching": {
          "batchSize": 1
        }
      },
      "id": "generate",
      "name": "Generate All Platform Content (Ollama)",
      "type": "@n8n/n8n-nodes-langchain.chainLlm",
      "typeVersion": 1.7,
      "position": [
        680,
        300
      ]
    },
    {
      "parameters": {
        "mode": "runOnceForAllItems",
        "jsCode": "const parse = function parseSocial({ text, topic, original = null }) {\n  const fields = ['twitter', 'linkedin', 'reddit_title', 'reddit_body', 'instagram_caption'];\n  let parsed;\n  try {\n    if (typeof text !== 'string') throw new Error('Model output must be text.');\n    const match = text.match(/\\{[\\s\\S]*\\}/);\n    if (!match) throw new Error('No JSON object found.');\n    parsed = JSON.parse(match[0]);\n    if (!parsed || Array.isArray(parsed) || fields.some((field) => typeof parsed[field] !== 'string' || !parsed[field].trim())) throw new Error('Model output must contain all five nonempty social content fields.');\n  } catch (error) {\n    if (original) return { result: { ...original, reviewed: false } };\n    throw new Error('Failed to parse AI response as complete content JSON: ' + error.message);\n  }\n  const content = Object.fromEntries(fields.map((field) => [field, parsed[field]]));\n  return { result: original ? { ...original, ...content, reviewed: true } : { ...content, topic, generated_at: new Date().toISOString() } };\n}\n\nreturn [{json: parse({text: $input.first().json.text, topic: $('Content Parameters').first().json.topic}).result}];\n"
      },
      "id": "parse",
      "name": "Parse Content JSON",
      "type": "n8n-nodes-base.code",
      "typeVersion": 2,
      "position": [
        900,
        300
      ]
    },
    {
      "parameters": {
        "promptType": "define",
        "text": "=Review this social media content for quality. Check for:\n1. Grammar and spelling errors\n2. Tone consistency with brand voice\n3. Platform-appropriate formatting\n4. Hashtag relevance\n\nBrand voice: {{ $('Content Parameters').first().json.brand_voice }}\nContent:\nTwitter: {{ $json.twitter }}\nLinkedIn: {{ $json.linkedin }}\nReddit: {{ $json.reddit_title }} — {{ $json.reddit_body }}\nInstagram: {{ $json.instagram_caption }}\n\nIf any issues found, return corrected versions as JSON (same structure). If all good, return the same content. Return ONLY JSON.",
        "batching": {
          "batchSize": 1
        }
      },
      "id": "review",
      "name": "Review & Polish (Ollama)",
      "type": "@n8n/n8n-nodes-langchain.chainLlm",
      "typeVersion": 1.7,
      "position": [
        1120,
        300
      ]
    },
    {
      "parameters": {
        "mode": "runOnceForAllItems",
        "jsCode": "const parse = function parseSocial({ text, topic, original = null }) {\n  const fields = ['twitter', 'linkedin', 'reddit_title', 'reddit_body', 'instagram_caption'];\n  let parsed;\n  try {\n    if (typeof text !== 'string') throw new Error('Model output must be text.');\n    const match = text.match(/\\{[\\s\\S]*\\}/);\n    if (!match) throw new Error('No JSON object found.');\n    parsed = JSON.parse(match[0]);\n    if (!parsed || Array.isArray(parsed) || fields.some((field) => typeof parsed[field] !== 'string' || !parsed[field].trim())) throw new Error('Model output must contain all five nonempty social content fields.');\n  } catch (error) {\n    if (original) return { result: { ...original, reviewed: false } };\n    throw new Error('Failed to parse AI response as complete content JSON: ' + error.message);\n  }\n  const content = Object.fromEntries(fields.map((field) => [field, parsed[field]]));\n  return { result: original ? { ...original, ...content, reviewed: true } : { ...content, topic, generated_at: new Date().toISOString() } };\n}\n\nreturn [{json: parse({text: $input.first().json.text, topic: $('Content Parameters').first().json.topic, original: $('Parse Content JSON').first().json}).result}];\n"
      },
      "id": "final-parse",
      "name": "Final Content Output",
      "type": "n8n-nodes-base.code",
      "typeVersion": 2,
      "position": [
        1340,
        300
      ]
    },
    {
      "id": "generate-ollama-model",
      "name": "Generate All Platform Content (Ollama) model",
      "type": "@n8n/n8n-nodes-langchain.lmOllama",
      "typeVersion": 1,
      "position": [
        680,
        500
      ],
      "parameters": {
        "model": "llama3:8b",
        "options": {
          "temperature": 0.7,
          "numPredict": 2000
        }
      }
    },
    {
      "id": "review-ollama-model",
      "name": "Review & Polish (Ollama) model",
      "type": "@n8n/n8n-nodes-langchain.lmOllama",
      "typeVersion": 1,
      "position": [
        1120,
        500
      ],
      "parameters": {
        "model": "llama3:8b",
        "options": {
          "temperature": 0.2,
          "numPredict": 2000
        }
      }
    }
  ],
  "connections": {
    "Manual Trigger": {
      "main": [
        [
          {
            "node": "Content Parameters",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Content Parameters": {
      "main": [
        [
          {
            "node": "Generate All Platform Content (Ollama)",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Generate All Platform Content (Ollama)": {
      "main": [
        [
          {
            "node": "Parse Content JSON",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Parse Content JSON": {
      "main": [
        [
          {
            "node": "Review & Polish (Ollama)",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Review & Polish (Ollama)": {
      "main": [
        [
          {
            "node": "Final Content Output",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Generate All Platform Content (Ollama) model": {
      "ai_languageModel": [
        [
          {
            "node": "Generate All Platform Content (Ollama)",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "Review & Polish (Ollama) model": {
      "ai_languageModel": [
        [
          {
            "node": "Review & Polish (Ollama)",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    }
  },
  "settings": {
    "executionOrder": "v1"
  },
  "tags": [
    {
      "name": "AI"
    },
    {
      "name": "Ollama"
    },
    {
      "name": "Social Media"
    },
    {
      "name": "Content"
    }
  ],
  "active": false,
  "pinData": {}
}