⚡ iMessage food photo nutritional analysis with GPT-4 Vision & Memory Storage

⚡ 2,073 views · ⚡ Personal Productivity

Description

iMessage AI-Powered Smart Calorie Tracker

> 📌 What it looks like in use:

> This image shows a visual of the workflow in action. Use it for reference when replicating or customizing the template.

This n8n template transforms a user-submitted food photo into a detailed, friendly, AI-generated nutritional report — sent back seamlessly as a chat message. It combines OpenAI’s visual reasoning, Postgres-based memory, and real-time messaging with Blooio to create a hands-free calorie and nutrition tracker.

🧠 Use Cases

• Auto-analyze meals based on user-uploaded images.
• Daily/weekly/monthly diet summaries with no manual input.
• Virtual food journaling integrated into messaging apps.
• Nutrition companion for healthcare, fitness, and wellness apps.

📌 Good to Know

• ⚠️ This uses GPT-4 with image capabilities, which may incur higher usage costs depending on your OpenAI pricing tier. Review OpenAI’s pricing.
• The model uses visual reasoning and estimation to determine nutritional info — results are estimates and should not replace medical advice.
• Blooio is used for sending/receiving messages. You will need a valid API key and project set up with webhook delivery.
• A Postgres database is required for long-term memory (optional but recommended). You can use any memory node with it.

⚙️ How It Works

1. Webhook Trigger
   The workflow begins when a message is received via Blooio. This webhook listens for user-submitted content, including any image attachments.

2. Image Validation and Extraction
   A conditional check verifies the presence of attachments. If images are found, their URLs are extracted using a Code node and prepared for processing.

3. Image Analysis via AI Agent
   Images are passed to an OpenAI-based agent using a custom system prompt that:

   • Identifies the meal,
   • Estimates portion sizes,
   • Calculates calories, macros, fiber, sugar, and sodium,
   • Scores the meal with a health and confidence rating,
   • Responds in a chatty, human-like summary format.

4. Memory Integration
   A Postgres memory node stores user interactions for recall and contextual continuity, allowing day/week/month reports to be generated based on cumulative messages.

5. Response Aggregation & Summary
   Messages are aggregated and summarized by a second AI agent into a single concise message to be sent back to the user via Blooio.

6. Message Dispatch
   The final message is posted back to the originating conversation using the Blooio Send Message API.

🚀 How to Use

• The included webhook can be triggered manually or programmatically by linking Blooio to a frontend chat UI.
• You can test the flow using a manual POST request containing mock Blooio payloads.
• Want to use a different messages app? Replace the Blooio nodes with your preferred messaging API (e.g., Twilio, Slack, Telegram).

✅ Requirements

• OpenAI API access with GPT-4 Vision or equivalent multimodal support.
• Blooio account with access to incoming and outgoing message APIs.
• Optional: Postgres DB (e.g., via Neon) for tracking message context over time.

🛠️ Customising This Workflow

• Prompt Tuning
  Tailor the system prompt in the AI Agent node to fit specific diets (e.g., keto, diabetic), age groups, or regionally-specific foods.

• Analytics Dashboards
  Hook up your Postgres memory to a data visualization tool for nutritional trends over time.

• Multilingual Support
  Adjust the response prompt to translate messages into other languages or regional dialects.

• Image Preprocessing
  Insert a preprocessing node before sending images to the model to resize, crop, or enhance clarity for better results.

🔗 Nodes Used

HTTP Request, Webhook, AI Agent, OpenAI Chat Model, Postgres Chat Memory

📥 Import

Download workflow.json and import into n8n: Workflow menu → Import from File

📖 Importing guide · 🔑 Credential setup