🤖 AutoDevOps AI

An AI-powered DevOps automation CLI — scan, deploy, diagnose, and document any project. Now live on PyPI.

---

📖 Table of Contents

• Overview
• What's New in v0.1.0
• Features
• Architecture
• Tech Stack
• Installation
• Initial Configuration
• CLI Commands
• Interactive Shell (REPL)
• Project Structure
• Phase-wise Roadmap
• Contributing
• License

---

🌟 Overview

AutoDevOps AI is an intelligent, multi-agent DevOps automation platform distributed as a global Python CLI. It scans your project, understands your tech stack, and autonomously:

• 🔍 Scans your repository to detect languages, frameworks, databases, and infrastructure
• 🧠 Plans deployment strategies using AI (multi-agent Planner)
• 🔗 Integrates deeply with GitHub (auth, repos, commits, PRs, secrets, releases)
• 🚀 Deploys to local environments, AWS, Azure, and GCP
• 🔧 Diagnoses CI/CD and deployment failures with root-cause AI analysis
• 📄 Generates full documentation, runbooks, and API docs
• 🛡️ Secures pipelines with secret detection and dependency scanning
• 💰 Optimizes cloud costs with actionable recommendations
• 🩺 Doctor mode scores your project across 5 health dimensions
• 🗺️ Roadmap mode generates a prioritized improvement plan

---

🆕 What's New in v0.1.1

• ✅ Published to PyPI — pip install autodevops-ai works globally
• ✅ Interactive REPL Shell — Claude/Gemini-style shell with slash commands, tab-completion, and persistent history
• ✅ autodevops doctor — project health scoring across Code, Security, Tests, Docs, and Deployment Readiness
• ✅ autodevops roadmap — AI-generated step-by-step improvement roadmap with time estimates
• ✅ autodevops undo — restore files from the last backup session
• ✅ autodevops run — multi-agent planner for high-level natural-language goals
• ✅ autodevops security — full secret detection, dependency scanning, .env auditing
• ✅ autodevops cost — cloud cost analysis and optimization recommendations
• ✅ GitHub sub-commands — commit, push, pull, pr, branch, release, review-pr, git-status, secrets
• ✅ Generate sub-commands — pipeline, dockerfile, code
• ✅ AI Providers — OpenAI, Anthropic, Google Gemini, Ollama (local) with exponential-backoff retry
• ✅ Core Engine — Chunker, Context, Memory, Streaming, Audit, Backup, Next-Actions

---

✨ Features

Feature	Description
🔍 Project Scanner	Detects languages, frameworks, Docker, K8s, Terraform, DBs, CI/CD, and cloud SDKs
🧩 Chunking Engine	Splits large repos into LLM-processable chunks with summaries
🧠 Context Engine	Builds rich AI context from scan + user request + chunks
💬 AI Memory	Maintains conversational context across shell session turns
🔗 GitHub Integration	Full GitHub automation: auth, repos, branches, commits, PRs, secrets, releases, reviews
🤖 AI Developer Assistant	Ask, explain, generate, refactor, and review code via CLI
📝 Smart Commit Generator	AI-generated commit messages, PR titles/descriptions, branch names, release notes
⚙️ CI/CD Generator	Generate pipelines for GitHub Actions, GitLab CI, Jenkins, Azure DevOps, CircleCI, Bitbucket
🖥️ Local Deployment	Deploy to Windows/Linux/macOS, Docker Desktop, Minikube
☁️ Cloud Deployment	Deploy to AWS (Lambda/ECS/EKS), Azure (App Service/AKS), GCP (Cloud Run/GKE)
🔧 Troubleshooting Agent	Diagnose failures from build, Docker, K8s, GitHub Actions, and cloud logs
📄 Documentation Agent	Generate README, architecture docs, API docs, runbooks
🛡️ Security Agent	Secret detection, dependency scanning, pipeline and .env file checks
💰 Cost Optimizer	Analyze and reduce AWS, Azure, and GCP resource costs
🩺 Doctor	Health score report across Code Quality, Security, Testing, Docs, Deployment Readiness
🗺️ Roadmap	Prioritized, actionable improvement roadmap with commands and time estimates
↩️ Undo	Rollback any AutoDevOps-generated file change via backup sessions
🖥️ Interactive REPL Shell	Persistent shell with slash commands, tab-completion, and AI chat

---

🏗️ Architecture

AutoDevOps AI follows a Multi-Agent Architecture where each agent has a clear responsibility:

+──────────────────────────────────────────────────────────────+
│                      AutoDevOps CLI                          │
│          (Typer-based, global `autodevops` command)          │
│        + Interactive REPL Shell (slash commands)             │
+──────────────────────────┬───────────────────────────────────+
                           │
               ┌───────────▼───────────┐
               │      Planner Agent    │
               │  (Multi-agent router) │
               └───────────┬───────────┘
                           │
    ┌──────────┬───────────┼───────────┬──────────┐
    │          │           │           │          │
┌───▼───┐ ┌───▼───┐ ┌─────▼────┐ ┌───▼───┐ ┌───▼────┐
│Scanner│ │Context│ │Developer │ │  Git  │ │CI/CD   │
│ Agent │ │Engine │ │  Agent   │ │ Agent │ │ Agent  │
└───────┘ └───────┘ └──────────┘ └───────┘ └────────┘
    │          │           │           │          │
┌───▼───┐ ┌───▼───┐ ┌─────▼────┐ ┌───▼───┐ ┌───▼────┐
│Deploy │ │Trouble│ │   Doc    │ │Securit│ │  Cost  │
│ Agent │ │-shoot │ │  Agent   │ │  y    │ │ Agent  │
└───────┘ └───────┘ └──────────┘ └───────┘ └────────┘
                           │
           ┌───────────────▼──────────────┐
           │   Core Engine                │
           │  Chunker · Context · Memory  │
           │  Streaming · Audit · Backup  │
           │  NextActions · Models        │
           └──────────────────────────────┘

Agent Responsibilities

Agent	Responsibility
Planner Agent	Orchestrates task decomposition, intent classification, and agent delegation
Scanner Agent	Rule-based project scanning — detects stack, infra, CI/CD, cloud SDKs (no AI)
Context Engine	Builds and optimizes AI context from scan results + chunked files
Developer Agent	Code generation, explanation, refactoring, review
Git Agent	Full GitHub API operations via PyGithub + gitpython
CI/CD Agent	Jinja2-based pipeline generation for 6 platforms
Deployment Agent	Local (native/Docker/Minikube) and cloud (AWS/Azure/GCP) automation
Documentation Agent	Generates README, architecture docs, API docs, runbooks
Troubleshooting Agent	Root-cause analysis and fix generation from logs
Security Agent	Secret detection, dependency scanning, pipeline and env file checks
Cost Optimization Agent	Cloud cost analysis (AWS/Azure/GCP) and right-sizing recommendations

---

🛠️ Tech Stack

Category	Libraries
CLI Framework	typer, rich
Interactive Shell	prompt_toolkit (REPL with autocomplete + history)
Data Validation	pydantic, pydantic-settings
AI / LLM	openai, anthropic, google-generativeai + Ollama (local)
Git / GitHub	gitpython, PyGithub
Containers	docker (Docker SDK)
Cloud (Optional)	boto3 (AWS), azure-mgmt-* + azure-identity, google-cloud-run
Kubernetes (Optional)	kubernetes
Templating	jinja2 (CI/CD pipeline templates)
Storage	tinydb (project memory/history)
HTTP	httpx, requests
File Analysis	pathspec
Token Counting	tiktoken
Secret Storage	keyring (OS keychain)
Config	python-dotenv
Testing	pytest, pytest-cov, pytest-asyncio
Packaging	setuptools, build, twine

---

📦 Installation

# Install globally from PyPI
pip install autodevops-ai

# Verify installation
autodevops --version

# Run the first-time setup wizard
autodevops setup

# Initialize in your project
autodevops init

---

⚙️ Initial Configuration

After installing, run the setup wizard or configure manually:

1. First-Time Setup Wizard (Recommended)

autodevops setup

Interactive wizard that configures AI provider keys, GitHub authentication, and project settings in one guided flow.

2. Set AI Provider API Key Manually

# OpenAI (Default)
autodevops config set openai_api_key sk-xxxxxxxxxxxxxxxxxxxx

# Anthropic Claude
autodevops config set anthropic_api_key sk-ant-xxxxxxxxxxxx
autodevops config set default_provider anthropic

# Google Gemini
autodevops config set google_api_key AIzaSy-xxxxxxxxxxxxxxxx
autodevops config set default_provider google

# Local Ollama (no API key needed)
autodevops config set default_provider ollama
autodevops config set ollama_model llama3

3. Connect GitHub

autodevops github connect
# Prompts for your GitHub Personal Access Token (PAT)
# Required scopes: repo, workflow, admin:repo_hook

---

💻 CLI Commands

Core Commands

autodevops                        # Launch the interactive REPL shell (default)
autodevops setup                  # First-time setup wizard
autodevops init [--name <n>]      # Initialize AutoDevOps in current project
autodevops scan [--output table|json|file] [--save]   # Scan project stack
autodevops ask "<question>"       # Ask AI about your project
autodevops explain [architecture|code|deployment]     # Explain project via AI
autodevops diagnose               # AI root-cause analysis of failures
autodevops docs [--type all|readme|api|runbook|architecture|deployment]  # Generate docs
autodevops config <show|set|reset> [key] [value]      # Manage configuration
autodevops status [--env local|aws|azure|gcp|all]     # Check deployment status
autodevops doctor                 # Project health score report
autodevops roadmap                # AI-generated improvement roadmap
autodevops security [--checks all|secrets|dependencies|pipeline|env_files]  # Security scan
autodevops cost [--cloud aws|azure|gcp|all] [--days 30]  # Cloud cost analysis
autodevops run "<goal>"           # Multi-agent planner for natural-language goals
autodevops undo                   # Restore files from last backup session
autodevops shell                  # Explicitly launch interactive shell
autodevops --version              # Show version

GitHub Sub-Commands

autodevops github connect [--token <pat>]                     # Connect to GitHub
autodevops github status                                       # Enhanced git status + AI tip
autodevops github commit [-m "<msg>"] [--no-push] [-b <branch>]  # AI smart commit
autodevops github push [-b <branch>] [--force] [-u]           # Push to remote
autodevops github pull [-b <branch>] [--rebase]               # Pull from remote
autodevops github pr [--title "<t>"] [--base main] [--draft]  # Create pull request
autodevops github branch [description]                         # AI branch name generator
autodevops github secrets <list|set|delete> [name] [value]    # Manage GitHub secrets
autodevops github release [--tag v1.2.0]                      # AI release notes + CHANGELOG
autodevops github review-pr [<number>]                         # AI code review of a PR

Deploy Sub-Commands

autodevops deploy local [--env development] [--docker]        # Deploy locally
autodevops deploy [--cloud aws|azure|gcp] [--service <svc>] [--dry-run]  # Cloud deploy

Generate Sub-Commands

autodevops generate pipeline [--platform github-actions|gitlab|jenkins|azure-devops|circleci|bitbucket]
autodevops generate dockerfile
autodevops generate code "<description>" [--lang <language>] [-o <output>]

autodevops run Examples (Multi-Agent Planner)

autodevops run "deploy my FastAPI app to AWS ECS"
autodevops run "generate a GitHub Actions pipeline and create a PR"
autodevops run "scan project, fix security issues, and generate documentation"
autodevops run "create a release v1.0.0 with changelog" --dry-run

---

🖥️ Interactive Shell (REPL)

Run autodevops (no arguments) or autodevops shell to enter the interactive REPL — a Claude CLI/Gemini CLI-style session.

Features:

• Slash command execution with tab-completion
• Persistent command history across sessions (~/.autodevops/shell_history)
• Plain text → streamed AI answer with full project context
• AI session memory (multi-turn conversation within the session)
• Arrow key history navigation
• Ctrl+C cancels the current line (does NOT exit)
• Ctrl+D or /exit exits cleanly

Available slash commands:

Command	Description
/help	Show all available commands
/init	Initialize AutoDevOps in this project
/scan	Scan project stack
/doctor	Project health score report
/roadmap	Generate improvement roadmap
/undo	Restore files from last backup
/ask <question>	Ask AI (same as plain text)
/commit	AI smart commit + push
/push	Push to remote
/pull	Pull from remote
/pr	Create a pull request
/branch [description]	AI branch name generator
/release [--tag v1.0.0]	AI release notes + CHANGELOG
/review-pr [number]	AI code review of a PR
/git-status	Enhanced git status with AI recommendation
/secrets	Manage GitHub secrets
/security	Run security scan
/docs	Generate documentation
/deploy	Deploy project
/diagnose	Diagnose failures
/cost	Analyze cloud costs
/generate	Generate files (pipeline, dockerfile, code)
/run <goal>	Multi-agent planner goal
/connect	Connect to GitHub
/status	Check deployment status
/config	Manage configuration
/history	Show recent shell history
/session clear	Clear AI conversation session
/clear	Clear the screen
/exit or /quit	Exit the AutoDevOps shell

---

📁 Project Structure

autodevops-ai/
├── src/
│   └── autodevops/
│       ├── __init__.py                 # Package version (0.1.0)
│       ├── cli.py                      # Main Typer CLI entry point (578 lines)
│       ├── config.py                   # ConfigManager — keys, keychain, dotenv
│       ├── agents/
│       │   ├── planner.py              # Multi-agent orchestrator + intent classifier
│       │   ├── scanner.py              # Rule-based project scanner (29 KB)
│       │   ├── cicd.py                 # CI/CD pipeline generator
│       │   ├── deployment.py           # Local + cloud deployment agent
│       │   ├── documentation.py        # Documentation generator agent
│       │   ├── troubleshoot.py         # Root-cause analysis agent
│       │   ├── security.py             # Security scanning agent (22 KB)
│       │   └── cost.py                 # Cloud cost analysis agent
│       ├── commands/
│       │   ├── ask.py                  # `autodevops ask` handler
│       │   ├── config.py               # `autodevops config` handler
│       │   ├── deploy.py               # `autodevops deploy` handler
│       │   ├── diagnose.py             # `autodevops diagnose` handler
│       │   ├── docs.py                 # `autodevops docs` handler
│       │   ├── doctor.py               # `autodevops doctor` health check
│       │   ├── explain.py              # `autodevops explain` handler
│       │   ├── generate.py             # `autodevops generate` handler
│       │   ├── github.py               # GitHub sub-command handlers
│       │   ├── handlers.py             # Shared command handler utilities
│       │   ├── init.py                 # `autodevops init` handler
│       │   ├── roadmap.py              # `autodevops roadmap` handler
│       │   ├── scan.py                 # `autodevops scan` handler
│       │   ├── setup.py                # `autodevops setup` wizard
│       │   ├── status.py               # `autodevops status` handler
│       │   └── undo.py                 # `autodevops undo` handler
│       ├── core/
│       │   ├── chunker.py              # LLM-safe file chunking engine
│       │   ├── context.py              # Context building + token budget management
│       │   ├── memory.py               # AI conversation memory (TinyDB)
│       │   ├── models.py               # Pydantic data models
│       │   ├── next_actions.py         # Next-action recommendation engine
│       │   ├── repl.py                 # Interactive REPL shell (734 lines)
│       │   ├── streaming.py            # Streaming AI response helper
│       │   ├── audit.py                # Operation audit trail
│       │   └── backup.py               # File backup + restore for undo
│       ├── providers/
│       │   └── base.py                 # Unified AI provider (OpenAI/Anthropic/Gemini/Ollama)
│       ├── integrations/
│       │   └── github.py               # Deep GitHub integration (45 KB, PyGithub + gitpython)
│       └── templates/
│           ├── github_actions/         # GitHub Actions pipeline templates
│           ├── gitlab_ci/              # GitLab CI templates
│           ├── jenkins/                # Jenkinsfile templates
│           ├── azure_devops/           # Azure Pipelines templates
│           ├── circleci/               # CircleCI config templates
│           ├── bitbucket/              # Bitbucket Pipelines templates
│           └── docs/                   # Documentation templates
├── tests/
│   ├── unit/
│   │   ├── test_scanner.py
│   │   ├── test_chunker.py
│   │   ├── test_planner.py
│   │   └── test_security.py
│   ├── integration/                    # (planned)
│   └── conftest.py
├── docs/                               # (planned — MkDocs)
├── pyproject.toml                      # Package metadata + build config
├── README.md
└── USER_GUIDE.md                       # Detailed user guide

🤝 Contributing

Contributions are welcome! Please read CONTRIBUTING.md for guidelines.

git clone https://github.com/akashhugar2015/AutoDevOps_AI.git
cd AutoDevOps_AI
pip install -e ".[dev]"
pytest tests/
ruff check src/

---

📄 License

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

---

Built with ❤️ to make DevOps intelligent, automated, and accessible to every developer.