one-click-ai-spark 0.3.1

One command to generate a production-ready AI/ML backend — LLM, RAG, Voice, Vision, Emotion, Agents, ML Training, Computer Vision, Edge AI & more

🤖 One Click AI Spark 🚀

One Click AI Spark is a powerful CLI tool that generates production-ready AI backend projects in seconds — powered by FastAPI, with built-in LLM integration, RAG pipelines, voice, vision, emotion detection, AI agents, ML training (PyTorch, TensorFlow, scikit-learn, XGBoost, LightGBM), computer vision (YOLO, SAM2, OCR, face detection, image generation), edge AI (ONNX, TensorRT, quantization), LLM fine-tuning (LoRA/QLoRA), AI guardrails (PII, content filter, prompt injection), conversational analytics (Text-to-SQL), Docker, CI/CD, IaC, and full observability.

🛠️ Technology Stack

⭐ Star us on GitHub — your support motivates us a lot! 🙏😊

---

⚡ Stop copy-pasting boilerplate. Start building what matters. ⚡

🌟 The Story Behind One Click AI Spark

Every time a new AI project starts, I see the same cycle — developers lose days reinventing the same infrastructure:

• Provider Headaches: Writing LLM client wrappers for OpenAI, Anthropic, Google, Groq — over and over.
• RAG Nightmares: Stitching together document ingestion, chunking, embedding, vector stores, and retrieval from scratch.
• Multimodal Chaos: Integrating STT, TTS, vision, and emotion detection with no standard structure.
• Infrastructure Fatigue: Manually setting up Dockerfiles, CI/CD pipelines, Terraform configs, Prometheus dashboards — before writing a single line of AI logic.
• Confidence Gap: After hours of wiring, nobody is 100% sure the architecture is production-ready.

One Click AI was born to eliminate this. It generates a battle-tested, modular AI backend in seconds — with every integration wired up, every best practice baked in, and every deployment tool ready to go. No more boilerplate, no more forgotten configs — just one command to start building real AI applications.

🚀 What You Get with ocd-ai

When you run ocd-ai init, you don't just get a folder — you get a complete AI production environment:

• 🧠 Multi-Provider LLM: Pre-configured clients for OpenAI, Anthropic, Google, Groq, Grok, Cohere, Mistral, and Ollama — swap providers with a config change.
• 📚 RAG Pipeline: Full document ingestion → smart chunking → embedding → vector retrieval → LLM generation pipeline, ready to go.
• 🎤 Voice & Audio: Whisper/Deepgram STT + OpenAI/ElevenLabs TTS + real-time voice-to-voice conversation.
• 👁️ Vision & Emotion: Image/video analysis (GPT-4o, Claude, Gemini) + emotion detection (Hume AI + LLM fallback).
• 🤖 AI Agents: ReAct-style agents with built-in tool use and web search (Tavily, Serper, DuckDuckGo).
• 🧠 Memory: Short-term (in-context) + long-term (persistent) conversation memory with session management.
• ⚡ Streaming: SSE + WebSocket streaming for real-time responses.
• 🐳 Docker Ready: Multi-stage Dockerfiles + docker-compose for dev and production with Nginx reverse proxy.
• ⚙️ CI/CD (GitHub Actions): Pre-configured workflows for testing, linting, and automated deployment.
• 🏗️ Infrastructure as Code: Terraform templates for AWS + Ansible playbooks for server configuration.
• 📊 Monitoring & Observability: Prometheus metrics + Grafana dashboards + alerting rules — all pre-configured.
• 📊 Vector Store Flexibility: Choose from FAISS, Pinecone, Qdrant, Weaviate, ChromaDB, Milvus, or pgvector.
• 🔒 Standard .gitignore & .env: Pre-configured environment files with secrets management best practices.
• 🚀 GitHub Integration: Auto-initializes a git repo and pushes to a new GitHub repository in one go.
• 🔬 ML Training Pipelines: PyTorch, TensorFlow, scikit-learn, XGBoost, LightGBM — with trainers, predictors, feature engineering, model registry, and experiment tracking (MLflow/W&B).
• 🖼️ Computer Vision: YOLO (object detection), SAM2 (segmentation), EasyOCR/Tesseract (OCR), face detection/recognition, Stable Diffusion (image generation).
• ⚡ Edge AI Deployment: ONNX Runtime, TensorRT conversion, INT8 quantization, model optimization, benchmarking.
• 🔧 LLM Fine-Tuning: LoRA/QLoRA fine-tuning with PEFT + Transformers, merge & export for deployment.
• 🛡️ AI Guardrails: Content safety (Detoxify), PII detection/anonymization (Presidio), prompt injection defense, audit logging.
• 📊 Conversational Analytics: Natural language → SQL (Text-to-SQL), auto-chart generation (Plotly), report builder.
• 🏠 Ollama Local Deployment: Self-hosted LLM inference via Ollama in docker-compose.
• 🚀 One-Command Deploy: Deploy to AWS ECS, Google Cloud Run, Azure Container Apps, DigitalOcean, Railway, Fly.io, Render, or any VPS — with ocd-ai deploy.

✨ Feature Matrix

Feature	Description
🧠 LLM Chat	Multi-provider support (OpenAI, Anthropic, Google, Groq, Grok, Cohere, Mistral, Ollama)
📚 RAG Pipeline	Document ingestion → chunking → embedding → retrieval → generation
🎤 Voice (STT/TTS)	Speech-to-Text (Whisper, Deepgram) + Text-to-Speech (OpenAI, ElevenLabs)
🗣️ Voice-to-Voice	Real-time audio conversation pipeline
👁️ Vision	Image & video analysis with GPT-4o, Claude, Gemini
😊 Emotion Detection	Audio/text emotion analysis (Hume AI + LLM fallback)
🔍 Web Search	Internet search (Tavily, Serper, DuckDuckGo) with LLM summarization
🤖 AI Agents	ReAct-style agents with tool use
🧠 Memory	Short-term + long-term conversation memory
⚡ Streaming	SSE + WebSocket streaming responses
🔐 Sessions	Persistent session management (Redis-backed)
📊 Vector Stores	FAISS, Pinecone, Qdrant, Weaviate, ChromaDB, Milvus, pgvector
🐳 Docker	Production + dev Dockerfiles, docker-compose, Nginx
🔄 CI/CD	GitHub Actions (test, lint, deploy)
🏗️ IaC	Terraform (AWS) + Ansible server configuration
📈 Monitoring	Prometheus + Grafana + alerts
🔬 ML Training	PyTorch, TensorFlow, scikit-learn, XGBoost, LightGBM pipelines
🖼️ Computer Vision	YOLO detection, SAM2 segmentation, OCR, face detection, image generation
⚡ Edge AI	ONNX Runtime, TensorRT, INT8 quantization, model benchmarking
🔧 Fine-Tuning	LoRA/QLoRA fine-tuning with PEFT + Transformers
🧪 MLOps	MLflow + Weights & Biases experiment tracking
🛡️ Guardrails	Content filter (Detoxify), PII detection (Presidio), prompt injection defense
📊 Analytics	Text-to-SQL, auto-chart generation (Plotly), report builder
🏠 Ollama	Local LLM deployment via docker-compose
🎯 GPU Support	CUDA-accelerated training, mixed precision (FP16/BF16)
📋 Audit Logging	Full AI interaction audit trail for compliance
🚀 One-Command Deploy	Deploy to AWS, GCP, Azure, DigitalOcean, Railway, Fly.io, Render, or VPS

📦 Installation

Since one-click-ai is a CLI tool, it is recommended to install it globally using uv:

uv tool install one-click-ai

If your system doesn't have uv yet, follow the installation guide here.

Alternatively, you can use pip:

pip install one-click-ai

🚀 Usage

Initialize a Full AI Project (everything enabled)

ocd-ai init my-ai-project --all

This creates a new folder my-ai-project with a complete production-ready AI backend.

Initialize in the Current Directory

ocd-ai init .

🤖 Chatbot (LLM Only)

ocd-ai init my-chatbot --llm-only --openai --docker

📚 RAG Application

ocd-ai init my-rag-app --rag --openai --faiss --docker --ci-cd

🎤 Voice Assistant

ocd-ai init my-voice-app --voice --voice-to-voice --openai --streaming --docker

👁️ Vision + Emotion Analysis

ocd-ai init my-vision-app --vision --emotion --openai --google --docker

🤖 AI Agent with Search & Memory

ocd-ai init my-agent --agents --search --memory --openai --streaming --docker --ci-cd

🚀 Full Multi-Provider Stack

ocd-ai init my-enterprise-app \
    --rag --agents --search --memory --streaming --voice --vision --emotion \
    --openai --anthropic --google --groq \
    --faiss --qdrant \
    --postgres --redis \
    --docker --ci-cd --iac --monitoring

🔬 ML Training Pipeline

ocd-ai init my-ml-project --ml-training --pytorch --sklearn --xgboost --mlops --docker

🖼️ Computer Vision

ocd-ai init my-cv-project --computer-vision --yolo --sam --ocr --face-detection --gpu --docker

🔧 LLM Fine-Tuning (LoRA/QLoRA)

ocd-ai init my-finetune --fine-tuning --pytorch --gpu --mlops --docker

⚡ Edge AI Deployment

ocd-ai init my-edge-project --edge-ai --onnx --tensorrt --quantization --docker

🛡️ AI Guardrails

ocd-ai init my-safe-ai --guardrails --openai --rag --docker

📊 Conversational Analytics

ocd-ai init my-analytics --analytics --openai --docker

🖼️ Image Generation

ocd-ai init my-image-gen --computer-vision --image-gen --gpu --docker

🚀 One-Command Deploy

Every generated project includes a deploy.py and deploy.sh that supports 8 deployment platforms out of the box:

# Deploy to any platform (interactive wizard if no platform specified)
ocd-ai deploy

# Direct deploy to a specific platform
ocd-ai deploy aws           # AWS ECS (Fargate)
ocd-ai deploy gcp           # Google Cloud Run
ocd-ai deploy azure         # Azure Container Apps
ocd-ai deploy digitalocean  # DigitalOcean App Platform
ocd-ai deploy railway       # Railway
ocd-ai deploy fly           # Fly.io
ocd-ai deploy render        # Render
ocd-ai deploy vps           # Any VPS via SSH + Docker

# Build & push Docker image before deploying
ocd-ai deploy aws --build --push

🎛️ CLI Flags

Feature Flags

Flag	Description
--all	Enable every feature, provider, vector store, database, and infrastructure option
--llm-only	Just LLM chat (no RAG, voice, vision, etc.)
--rag	RAG pipeline (ingestion, chunking, embedding, retrieval, generation)
--voice	Speech-to-Text + Text-to-Speech
--voice-to-voice	Voice conversation (implies --voice + --streaming)
--vision	Image/video analysis
--emotion	Emotion detection (audio + text)
--search	Web search with LLM summarization
--agents	ReAct-style AI agents with tool use
--memory	Short-term + long-term conversation memory
--streaming	SSE + WebSocket streaming
--session	Redis-backed session management

LLM Providers

Flag	Provider
--openai	OpenAI (GPT-4o, GPT-4, GPT-3.5)
--anthropic	Anthropic (Claude 3.5, Claude 3)
--google	Google (Gemini Pro, Gemini Flash)
--groq	Groq (Llama, Mixtral — ultra-fast inference)
--grok	xAI Grok
--cohere	Cohere (Command R+)
--mistral	Mistral AI (Mistral Large, Mixtral)
--ollama	Ollama (local models — Llama, Mistral, etc.)
--all-providers	Enable all providers

Vector Stores

Flag	Store
--faiss	Facebook AI Similarity Search (local, fast)
--pinecone	Pinecone (managed, scalable)
--qdrant	Qdrant (open-source, feature-rich)
--weaviate	Weaviate (hybrid search)
--chroma	ChromaDB (lightweight, embedded)
--milvus	Milvus (distributed, GPU-accelerated)
--pgvector	pgvector (PostgreSQL extension)

Databases

Flag	Database
--postgres	PostgreSQL (primary relational database)
--mongodb	MongoDB (document store)
--redis	Redis (caching, sessions, pub/sub)
--all-databases	Enable all databases

Infrastructure

Flag	Description
--docker	Dockerfiles + docker-compose (dev & prod) + Nginx
--ci-cd	GitHub Actions workflows (test, lint, deploy)
--iac	Terraform (AWS) + Ansible server configuration
--monitoring	Prometheus + Grafana + alerting rules

ML / Training

Flag	Description
--ml-training	Full ML training pipeline (trainer, predictor, data loader, model registry)
--pytorch	PyTorch models (MLP, CNN, LSTM, Transformer)
--tensorflow	TensorFlow/Keras models
--sklearn	scikit-learn pipelines (Random Forest, SVM, Gradient Boosting, auto-select)
--xgboost	XGBoost gradient boosting
--lightgbm	LightGBM gradient boosting
--fine-tuning	LLM fine-tuning with LoRA/QLoRA (PEFT + Transformers)
--gpu	Enable CUDA/GPU, mixed precision (FP16/BF16)
--mlops	MLflow + Weights & Biases experiment tracking

Computer Vision

Flag	Description
--computer-vision	Computer vision module (inference, pre/post-processing)
--yolo	YOLOv8/v11 object detection (Ultralytics)
--sam	SAM2 image segmentation (Segment Anything)
--ocr	OCR text extraction (EasyOCR + Tesseract)
--face-detection	Face detection & recognition (face-recognition + MediaPipe)
--image-gen	Text-to-image generation (Stable Diffusion / Diffusers)

Edge AI

Flag	Description
--edge-ai	Edge AI deployment module
--onnx	ONNX model conversion + ONNX Runtime inference
--tensorrt	TensorRT model optimization (NVIDIA)
--quantization	INT8 quantization (ONNX + HuggingFace Optimum)

Advanced Features

Flag	Description
--aggregator	AI API aggregator pattern (multi-provider routing)
--analytics	Conversational analytics (Text-to-SQL + charts + reports)
--guardrails	AI safety (content filter, PII detection, prompt injection defense, audit)
--multi-tenant	Multi-tenant architecture support
--ab-testing	A/B testing framework for prompts/models
--ollama-serve	Local Ollama deployment in docker-compose

📁 Generated Project Structure

my-ai-project/
├── backend/
│   ├── app/
│   │   ├── main.py                 # FastAPI application entry point
│   │   ├── config.py               # Pydantic settings & environment config
│   │   ├── dependencies.py         # Dependency injection
│   │   ├── exceptions.py           # Custom exception handlers
│   │   ├── api/v1/                 # API routes (versioned)
│   │   │   ├── health.py           # Health check endpoint
│   │   │   ├── chat.py             # LLM chat endpoint
│   │   │   ├── rag.py              # RAG query endpoint
│   │   │   ├── documents.py        # Document upload/management
│   │   │   ├── voice.py            # STT/TTS endpoints
│   │   │   ├── websocket.py        # WebSocket streaming
│   │   │   ├── vision.py           # Image/video analysis
│   │   │   ├── emotion.py          # Emotion detection
│   │   │   ├── search.py           # Web search
│   │   │   ├── agents.py           # AI agent endpoint
│   │   │   ├── sessions.py         # Session management
│   │   │   ├── ml.py               # ML training & prediction
│   │   │   ├── cv.py               # Computer vision endpoints
│   │   │   ├── edge.py             # Edge AI (convert, benchmark)
│   │   │   ├── analytics.py        # Text-to-SQL & charts
│   │   │   └── guardrails.py       # AI safety endpoints
│   │   ├── core/
│   │   │   ├── ai/                 # LLM clients + provider factory
│   │   │   │   ├── providers/      # OpenAI, Anthropic, Google, etc.
│   │   │   │   ├── llm_client.py   # Base LLM interface
│   │   │   │   └── embeddings.py   # Embedding generation
│   │   │   ├── rag/                # RAG pipeline
│   │   │   │   ├── ingestion.py    # Document ingestion
│   │   │   │   ├── chunking.py     # Smart text chunking
│   │   │   │   ├── retriever.py    # Vector retrieval
│   │   │   │   └── pipeline.py     # End-to-end RAG
│   │   │   ├── multimodal/         # Voice, vision, emotion
│   │   │   │   ├── stt.py          # Speech-to-Text
│   │   │   │   ├── tts.py          # Text-to-Speech
│   │   │   │   ├── vision.py       # Vision analysis
│   │   │   │   └── emotion.py      # Emotion detection
│   │   │   ├── agents/             # Agent framework
│   │   │   ├── memory/             # Memory management
│   │   │   └── search/             # Web search
│   │   ├── ml/                     # ML training pipeline
│   │   │   ├── trainer.py          # Unified trainer (PyTorch/sklearn/XGBoost)
│   │   │   ├── predictor.py        # Model inference
│   │   │   ├── data_loader.py      # Data loading & splitting
│   │   │   ├── feature_engineering.py # Feature scaling, encoding
│   │   │   ├── model_registry.py   # Model versioning & promotion
│   │   │   ├── evaluation.py       # Metrics (accuracy, F1, RMSE)
│   │   │   ├── experiment.py       # MLflow/W&B tracking
│   │   │   ├── pytorch_models.py   # MLP, CNN, LSTM, Transformer
│   │   │   ├── sklearn_models.py   # RF, SVM, GB, auto-select
│   │   │   ├── tf_models.py        # Keras models
│   │   │   └── fine_tuning.py      # LoRA/QLoRA fine-tuning
│   │   ├── cv/                     # Computer vision
│   │   │   ├── inference.py        # Unified CV pipeline
│   │   │   ├── preprocessing.py    # Resize, normalize, augment
│   │   │   ├── postprocessing.py   # NMS, drawing, formatting
│   │   │   ├── yolo_detector.py    # YOLOv8/v11 detection
│   │   │   ├── sam_segmenter.py    # SAM2 segmentation
│   │   │   ├── ocr_engine.py       # EasyOCR + Tesseract
│   │   │   ├── face_detector.py    # Face detection/recognition
│   │   │   └── image_generator.py  # Stable Diffusion generation
│   │   ├── edge/                   # Edge AI deployment
│   │   │   ├── converter.py        # PyTorch→ONNX→TensorRT
│   │   │   ├── optimizer.py        # Quantization, pruning
│   │   │   └── runtime.py          # ONNX Runtime inference
│   │   ├── analytics/              # Conversational analytics
│   │   │   ├── text_to_sql.py      # NL→SQL engine
│   │   │   ├── chart_gen.py        # Plotly chart generation
│   │   │   └── report.py           # HTML report builder
│   │   ├── guardrails/             # AI safety
│   │   │   ├── content_filter.py   # Toxicity detection (Detoxify)
│   │   │   ├── pii_detector.py     # PII detection (Presidio)
│   │   │   ├── prompt_injection.py # Injection defense
│   │   │   └── audit_logger.py     # Compliance audit trail
│   │   ├── services/               # Business logic layer
│   │   ├── models/                 # Pydantic schemas & enums
│   │   ├── db/                     # Database connections
│   │   └── utils/                  # Logger, security, helpers
│   ├── tests/                      # Test suite
│   ├── vector_stores/              # Vector store implementations
│   ├── middleware/                  # Error handling, rate limiting, logging
│   ├── Dockerfile                  # Production Dockerfile
│   ├── Dockerfile.dev              # Development Dockerfile
│   ├── .env                        # Environment variables
│   └── .env.example                # Environment template
├── docker-compose.dev.yml          # Development orchestration
├── docker-compose.prod.yml         # Production orchestration
├── nginx/
│   └── nginx.conf                  # Reverse proxy configuration
├── .github/
│   └── workflows/
│       ├── test.yml                # CI testing
│       ├── lint.yml                # Code linting
│       └── deploy.yml              # Deployment pipeline
├── infra/
│   ├── terraform/                  # AWS infrastructure
│   └── ansible/                    # Server configuration
├── monitoring/
│   ├── prometheus.yml              # Metrics collection
│   ├── alerts.yml                  # Alert rules
│   └── grafana_dashboard.json      # Pre-built dashboards
├── scripts/
│   ├── start.sh                    # Application launcher
│   └── health_check.py             # Health check script
├── models/                         # Trained model storage
│   ├── checkpoints/                # Training checkpoints
│   └── exported/                   # Production-ready models
├── datasets/                       # Training data
│   ├── raw/                        # Raw datasets
│   └── processed/                  # Processed features
├── experiments/                    # MLflow/W&B experiment logs
├── Makefile                        # Common commands
├── deploy.py                       # Cross-platform deploy script (Python)
├── deploy.sh                       # Deploy script (Bash)
├── .gitignore                      # Git ignore rules
└── DEVELOPMENT_GUIDE.md            # Getting started guide

🚀 Built for the Future (Extensibility)

This structure is intentionally designed for scaling into a full-stack or microservices architecture:

• Full-Stack Ready: Need a frontend? Add a frontend/ folder (React, Next.js, Streamlit, Gradio) at the root.
• Microservices Ready: Plug in other services like ml-training/, data-pipeline/, or notification-service/ alongside the backend.
• Multi-Model Ready: Each AI provider is isolated behind a factory pattern — adding a new provider is one file.
• ML/CV Pipeline Ready: Training, inference, and edge deployment modules are fully isolated and can scale independently.
• Simplified Orchestration: New services integrate into the root docker-compose.yml and CI/CD pipelines seamlessly.
• Guardrails by Default: Content safety, PII protection, and prompt injection defense can be applied to any endpoint.

🔧 Tech Stack

Category	Technologies
Framework	FastAPI ≥ 0.115
Runtime	Python ≥ 3.11, uvicorn
Package Manager	uv
Validation	Pydantic v2, pydantic-settings
LLM Providers	OpenAI, Anthropic, Google, Groq, Grok, Cohere, Mistral, Ollama, DeepSeek, Meta
Vector Stores	FAISS, Pinecone, Qdrant, Weaviate, ChromaDB, Milvus, pgvector, LanceDB
Voice	Whisper, Deepgram, whisper-local (STT) · OpenAI TTS, ElevenLabs, Bark (TTS)
Vision	GPT-4.1, Claude Opus 4, Gemini 2.5 Pro Vision
Emotion	Hume AI, Affectiva, LLM-based fallback
Search	Tavily, Serper, DuckDuckGo
ML Frameworks	PyTorch, TensorFlow/Keras, scikit-learn, XGBoost, LightGBM
Computer Vision	YOLO (Ultralytics), SAM2, EasyOCR, Tesseract, face-recognition, MediaPipe, Diffusers
Edge AI	ONNX, ONNX Runtime, TensorRT, HuggingFace Optimum
Fine-Tuning	PEFT (LoRA/QLoRA), Transformers, TRL, bitsandbytes
MLOps	MLflow, Weights & Biases
Guardrails	Detoxify, Presidio (PII), custom prompt injection defense
Analytics	SQLAlchemy, sqlparse, Plotly
Databases	PostgreSQL, MongoDB, Redis
Containers	Docker + docker-compose + Nginx
CI/CD	GitHub Actions
IaC	Terraform (AWS) + Ansible
Monitoring	Prometheus + Grafana + Alertmanager

⚙️ Configuration

On first run, ocd-ai will ask for your GitHub username, DockerHub username, default LLM provider, and default vector store. These are saved in ~/.config/one-click-ai/config.toml so you never have to type them again.

You can reset your configuration at any time by editing or deleting the config file.

🤝 Contributing

We welcome contributions! Here's how to get started:

1. Fork the repository
2. Clone your fork: git clone https://github.com/your-username/one-click-ai.git
3. Install dependencies: uv sync or pip install -e .
4. Create a branch: git checkout -b feature/my-feature
5. Make your changes — add generators in src/one_click_ai/generator.py or new templates in src/one_click_ai/templates/
6. Test locally: ocd-ai init test-project --all
7. Submit a Pull Request 🎉

See CONTRIBUTING.md for detailed guidelines.

👨‍💻 Author

Mohammad Asif Khan

• GitHub: @aktonay
• LinkedIn: Mohammad Asif Khan
• Website: sparktech.agency

📄 License

This project is licensed under the MIT License — see the LICENSE file for details.

---

Made with ❤️ for the AI developer community
Back to Top ⬆