ClickML is a low-code/no-code platform that helps MLOps engineers and data teams to create end-to-end ML pipelines — from ETL to model training and deployment — all through a simple, click-based interface.

Architecture

Workflow in ClickML

flowchart TB
    %% ── Storage Layer (top-left) ──────────────────────────────
    subgraph STORAGE["Storage Layer"]
        LogFiles[("Log Files\nPostgreSQL")]
        ModelDB[("Model File\nDatabase (S3)")]
        PlatformDB[("Platform-dependent\nDatabase (PostgreSQL)")]
    end

    subgraph USER_STORE["User Storage"]
        UserData[("User Data")]
        UserDatabase[("User Database")]
    end

    %% ── Auth ──────────────────────────────────────────────────
    UserA(["User"])
    UserB(["User"])
    SignUp["Sign Up"]
    Login["Login"]

    UserA --> SignUp --> UserDatabase
    UserB --> Login --> UserDatabase
    UserDatabase --> |"Authenticate"| InteractionLayer

    %% ── ETL Pipeline (top-center) ─────────────────────────────
    subgraph ETL["Airflow – ETL Pipeline"]
        RunDAG["Run DAG"]
        Extract["Extract Data"]
        Transform["Transform"]
        Load["Load Data"]
        Trigger1{{"Trigger"}}

        RunDAG --> Extract --> Transform --> Load --> Trigger1
    end

    APIConfig["API's State\nEndpoint / Secret Key"] --> ETL
    ETL --> |"Logs"| LogFiles
    Trigger1 --> |"Database?"| DBCheck{{"DB?"}}
    DBCheck --> |"Yes"| DataLake
    DBCheck --> |"No – Fetch data"| DataLake

    DataLake[("Data Lake /\nWarehouse")]

    %% ── Interaction Layer ─────────────────────────────────────
    subgraph InteractionLayer["Interaction Layer"]
        direction TB
        PipelineCreate["Data Pipeline Creation"]
        MLPipeline["ML / DL Pipeline"]
        ModelDeploy["Model Deployment"]
    end

    PipelineCreate --> |"Format (optional)\nTransform data format"| ETL
    PipelineCreate --> |"Database: hostname,\npassword, dbname"| ETL
    PipelineCreate --> |"Trigger Time"| ETL

    MLPipeline --> ModelSelection["Model Selection"]
    ModelDeploy --> Redeploy["Redeploy"]
    ModelDeploy --> ModelFileSelection["Model File Selection"]

    %% ── ML Training Pipeline (right) ─────────────────────────
    subgraph TRAINING["ML Training Pipeline"]
        direction TB
        Trigger2{{"Trigger"}}
        DataPreprocess["Data Preprocess"]
        ModelTrain["Model Train"]
        Evaluation["Evaluation"]
        TestVal["Test Validation"]

        Trigger2 --> DataPreprocess --> ModelTrain --> TestVal --> Evaluation
    end

    ModelSelection --> |"Model Type"| TRAINING
    ModelSelection --> |"Parameters"| TRAINING
    ModelSelection --> |"Input/Output Features"| TRAINING
    DataLake --> |"Fetch Data"| TRAINING

    Evaluation --> ModelReport["Model Report"]
    Evaluation --> ModelPKL[("Model\n(.pkl) File")]
    TRAINING --> |"Logs"| TrainingLogs[("Model Training\nLogs")]
    ModelPKL --> |"Storing output files"| DataLake

    %% ── Deployment (bottom-center) ────────────────────────────
    ModelFileSelection --> |"Model File (.pkl)"| DeployFlow

    subgraph DeployFlow["Deployment Flow"]
        FastAPI["Create Fast API Server"]
        EC2Deploy["Deploy on AWS EC2"]
        OutputJob["Output – Server Job"]

        FastAPI --> EC2Deploy --> OutputJob
    end

    DeployFlow --> |"Logs"| LogFiles

    %% ── Infrastructure (bottom-left) ──────────────────────────
    subgraph INFRA["Infrastructure (AWS)"]
        Terminal[">_ Terminal\nssh -i print-key clickml@ec2-ip"]
        EC2["EC2 Instance"]
        RDS[("RDS")]
        ClickMLDB[("clickml-database")]

        Terminal --> |"connect@ssh username"| EC2
        EC2 --> |"Insert username\n+ password"| RDS
        RDS --> ClickMLDB
    end

    InteractionLayer --> |"Send Models"| ModelSelection
    ModelPKL --> ModelFileSelection
    STORAGE --> InteractionLayer

    %% ── Styles ────────────────────────────────────────────────
    classDef storage fill:#4a90d9,stroke:#2c5f8a,color:#fff
    classDef process fill:#f9f3d9,stroke:#c8a84b,color:#333
    classDef decision fill:#ffe0b2,stroke:#e65100,color:#333
    classDef infra fill:#e8f5e9,stroke:#388e3c,color:#333
    classDef io fill:#fce4ec,stroke:#c62828,color:#333

    class LogFiles,ModelDB,PlatformDB,UserData,UserDatabase,DataLake,TrainingLogs,ModelPKL,ClickMLDB,RDS storage
    class Extract,Transform,Load,RunDAG,DataPreprocess,ModelTrain,Evaluation,FastAPI,EC2Deploy,ModelSelection process
    class Trigger1,Trigger2,DBCheck decision
    class EC2,Terminal infra
    class ModelReport,OutputJob io

Loading

---

Use Cases

Use Case	Description
Data Pipeline Automation	Connect to any database and build ETL pipelines without writing Airflow DAGs manually
No-Code Model Training	Select your dataset, choose an ML/DL algorithm, configure hyperparameters — and train with one click
One-Click Deployment	Deploy trained .pkl models as live FastAPI endpoints on AWS EC2 instantly
Model Monitoring	Track training logs, evaluation metrics, and model reports in one place
Redeployment	Swap out model versions and redeploy without touching infrastructure

Tip

For faster training, connect a Data Lake/Warehouse instead of a raw database — it reduces ETL overhead significantly.

Features

• ETL Pipeline Builder — Configure database credentials, set triggers, and run Airflow DAGs through the UI
• ML/DL Pipeline — Select model type, input/output features, and parameters visually
• Model Training — Automated data preprocessing → training → test validation → evaluation
• Model Deployment — Auto-generates a FastAPI server and deploys to AWS EC2
• Log Viewer — Real-time logs for ETL runs, training jobs, and deployed servers
• Model File Management — Download or redeploy stored .pkl model files from S3
• User Authentication — Secure sign-up/login with role-based access

Project Structure

ClickML/
├── Frontend/clickml/        # React/Next.js frontend (TypeScript)
├── Backend/                 # Core API server (authentication, pipeline management)
├── ETL-Backend/             # Airflow DAG generator and ETL orchestration
├── ML-Backend/              # Model training, evaluation, and serialization
├── airflow-processor/       # Airflow DAG definitions and processors
└── README.md

Installation & Setup

Prerequisites

• Python 3.9+
• Node.js 18+
• PostgreSQL
• Apache Airflow
• AWS account (EC2 + S3 + RDS)

Note

ClickML requires active AWS credentials to enable model deployment. Without them, training still works locally.

1. Clone the Repository

git clone https://github.com/AnantaXe/ClickML.git
cd ClickML

2. Frontend Setup

cd Frontend/clickml
npm install
npm run dev

3. Backend Setup

cd Backend
pip install -r requirements.txt
uvicorn main:app --reload

4. ML Backend Setup

cd ML-Backend
pip install -r requirements.txt
uvicorn main:app --reload --port 8001

5. ETL Backend + Airflow Setup

cd ETL-Backend
pip install -r requirements.txt

# Initialize Airflow
airflow db init
airflow webserver --port 8080 &
airflow scheduler &

Important

Make sure all 3 backends (Backend, ML-Backend, ETL-Backend) are running simultaneously for the full platform to work.

6. Environment Variables

Create a .env file in each service directory:

# Database
DB_HOST=your-db-host
DB_PORT=5432
DB_NAME=clickml
DB_USER=your-user
DB_PASSWORD=your-password

# AWS
AWS_ACCESS_KEY_ID=your-key
AWS_SECRET_ACCESS_KEY=your-secret
AWS_REGION=us-east-1
S3_BUCKET=your-bucket

# Airflow
AIRFLOW__CORE__EXECUTOR=LocalExecutor

Warning

Do not expose your .env file — it contains AWS keys and database credentials. Add it to .gitignore before pushing.

How to Use

Step 1 — Create an Account

Sign up at the platform URL and log in to access your dashboard.

Step 2 — Set Up a Data Pipeline

1. Go to Data Pipeline Creation
2. Enter your database credentials (hostname, port, database name, password)
3. Configure your data format and transformation rules
4. Set a trigger time or run manually
5. ClickML will auto-generate and deploy an Airflow DAG

Step 3 — Train a Model

1. Navigate to ML / DL Pipeline
2. Select your dataset (from the connected data source or Data Lake)
3. Choose a model type (classification, regression, etc.)
4. Define input/output features and set hyperparameters
5. Click Train — the platform handles preprocessing, training, validation, and evaluation

Step 4 — Deploy Your Model

1. Go to Model Deployment
2. Select a trained .pkl model file
3. Click Deploy — a FastAPI server is spun up on AWS EC2 automatically
4. Copy your live API endpoint and start sending inference requests

Step 5 — Monitor & Redeploy

• View real-time logs from the Logs panel
• Download model reports and evaluation metrics
• Use Redeploy to update a running endpoint with a new model version

Caution

Redeploying a model will replace the existing live endpoint. Ensure the new model is validated before redeployment.

Live Demo

Try the live demo at: click-ml-53za.vercel.app

Contributing

Contributions are welcome! Here's how to get started:

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

Please open an issue first for major changes to discuss what you'd like to change.