CalorieMate

CalorieMate is a smart food calorie and protein tracking application based on AI visual recognition technology. By simply taking photos, you can get nutritional estimates of your food, helping you easily manage your dietary intake. ###Core characteristics - ** One-click photo analysis ** -You can take a photo and analyze it without manual input - **AI intelligent recognition ** -Automatic recognition of food based on advanced visual AI models - ** Similar meal matching ** - CLIP image embedding technology to automatically match historical similar meals - ** Nutrition Tracking ** -Tracking daily calorie and protein intake - ** Goal Management ** -Set personalized nutrition goals - ** Data visualization ** -Daily/weekly calorie and protein statistics chart ###Application scenarios - Weight loss and shaping-control calorie intake - Fitness and muscle enhancement-Track protein intake - Health Management-Develop good eating habits - Diet Record-Create a personal diet file ###Step 1: Configure AI services CalorieMate needs AI services to analyze food photos. You have two choices: ####Option A: Use OpenRouter (recommended, simple and fast) OpenRouter is an AI API aggregation platform that can access multiple AI models (including Gemini, GPT-4V, etc.). 1. ** Get API Key** - Visit [OpenRouter official website](openrouter.ai) - Click "Sign In" in the upper right corner to register an account - Go to the "Keys" page and click "Create Key" - Copy your API Key (format: `sk-or-v1-xxxxx`) 2. ** Configure environment variables **(in Lazy Cat Micro Service) - Enter application settings - Add environment variables: ``` AI_PROVIDER=openrouter OPENROUTER_API_KEY=sk-or-v1-Your key OPENROUTER_VISION_MODEL=google/gemini-2.5-flash ``` - restart application 3. ** Cost description ** - OpenRouter Pay-Per-Use - Gemini 2.5 Flash model cost: approximately $0.0001/image - New users usually have free credit ####Option B: Use Olama (completely free, run locally) Olama is a local AI running tool that is completely free but requires better hardware. 1. ** Install Ollama** - macOS: `brew install ollama` - Windows: Download and install from [ ollama.ai](https://ollama.ai) - Linux: `curl -fsSL https://ollama.ai/install.sh | sh` 2. ** Launch Ollama service ** ```bash ollama serve ``` 3. ** Download visual model ** ```bash #Recommendation: 7B parametric model (about 4GB) ollama pull llava:7b #or larger model (more accurate, requires 8GB+ memory) ollama pull llava:13b ``` 4. ** Configure environment variables ** ``` AI_PROVIDER=ollama OLLAMA_HOST=http://host.docker.internal:11434 OLLAMA_VISION_MODEL=llava:7b ```