dockrion 0.0.2

Dockrion - Deploy and manage AI agents with ease

Dockrion

Deploy and manage AI agents with ease.

Dockrion provides a complete toolkit for building, validating, and deploying AI agents powered by LangGraph, LangChain, and other frameworks. Define your agent in a simple YAML file and deploy it anywhere.

🚀 Installation

pip install dockrion

Optional Features

# LangGraph framework support
pip install dockrion[langgraph]

# LangChain framework support
pip install dockrion[langchain]

# Runtime server (FastAPI + Uvicorn)
pip install dockrion[runtime]

# JWT authentication
pip install dockrion[jwt]

# Everything included
pip install dockrion[all]

⚡ Quick Start

1. Initialize a New Project

dockrion init my-agent
cd my-agent

This creates a Dockfile.yaml template:

version: "1.0"

agent:
  name: my-agent
  entrypoint: app.graph:build_agent
  framework: langgraph

io_schema:
  input:
    type: object
    properties:
      messages:
        type: array
  output:
    type: object
    properties:
      response:
        type: string

expose:
  port: 8080

2. Validate Your Configuration

dockrion validate

3. Test Locally

dockrion test --payload '{"messages": [{"role": "user", "content": "Hello!"}]}'

4. Run the Server

dockrion run

Your agent is now accessible at http://localhost:8080.

5. Build & Deploy

# Build Docker image
dockrion build

# Deploy to your infrastructure
dockrion deploy

📦 What's Included

The dockrion package bundles everything you need:

Module	Description
dockrion_cli	Command-line interface for all operations
dockrion_sdk	SDK for building and testing agents
dockrion_runtime	FastAPI server for deployed agents
dockrion_schema	Dockfile YAML schema validation
dockrion_adapters	Framework adapters (LangGraph, LangChain)
dockrion_policy	Policy engine for redaction & tool gating
dockrion_telemetry	Prometheus metrics & structured logging
dockrion_common	Shared utilities and error handling

✨ Features

• 🚀 Simple Configuration — Define your agent in a declarative YAML Dockfile
• 🔧 Framework Agnostic — Works with LangGraph, LangChain, and custom frameworks
• 🛡️ Built-in Safety — Policy engine for PII redaction and tool gating
• 📊 Observability — Prometheus metrics and structured JSON logging
• 🔐 Security — JWT authentication and API key support out of the box
• 🐳 Container Ready — Docker-first deployment with auto-generated Dockerfiles
• ⚡ Fast Development — Hot reload and local testing without containers
• 🧪 Testing Tools — Test your agent with payloads before deployment

🔧 CLI Commands

Command	Description
dockrion init <name>	Create a new agent project
dockrion validate	Validate Dockfile configuration
dockrion test	Test agent with sample payloads
dockrion run	Run agent server locally
dockrion build	Build Docker image
dockrion deploy	Deploy to target environment
dockrion logs	View agent logs
dockrion doctor	Diagnose setup issues
dockrion version	Show version information

📚 Documentation

For full documentation, examples, and guides:

• 📖 GitHub Repository
• 📋 Installation Guide
• 🏗️ Architecture

🤝 Contributing

We welcome contributions! Please see our GitHub repository for:

• Issue tracking
• Pull request guidelines
• Development setup instructions

👥 Authors

• Paritosh Sharma — [e-mail]
• Prakhar Agarwal — [e-mail]

📄 License

Apache-2.0 — See LICENSE for details.