Natural language → Code generation → Execution → Auto-debugging → Deployment.
All from a single CLI. No IDE needed.
npm install -g devos-ai
DevOS replaces your entire development workflow with a single intelligent CLI.
Describe what you want in plain English. DevOS plans, codes, tests, and deploys — fully autonomous.
devos goal "Build a Twitter clone with auth and deploy it"
Run → Detect errors → AI fixes → Re-run. Automatic until it works.
devos auto
Natural language terminal. Say "kill port 3000" instead of remembering commands.
devos shell
Auto-detects your stack and deploys to Vercel, Railway, Docker, Fly.io, or localhost.
devos deploy
Generates unit tests, runs them, and auto-fixes failures.
devos test --fix
Finds exposed API keys, SQL injection risks, XSS vulnerabilities, and more.
devos security
Intelligent code improvements — performance, cleanup, deduplication.
devos refactor
Monitors file changes and auto-fixes issues in real-time.
devos watch
Server crash? Auto-restart. Production bug? Auto-patch.
devos monitor --auto-fix
JS→TS, React→Vue, framework upgrades — automated transformations.
devos migrate typescript
Find any function, class, or pattern in your codebase with natural language.
devos find "authentication logic"
Track API usage and estimated costs across all providers.
devos cost
Six specialized AI agents collaborate to handle any development task.
Breaks requirements into actionable tasks with dependencies
Writes production-ready code, file by file
Runs commands, manages processes, detects errors
Analyzes errors, generates precise fixes
Quality checks, refactoring, security analysis
Builds, deploys, monitors in production
Coordinates all agents, manages task flow, handles retries and error recovery
| Command | Description | Example |
|---|---|---|
devos init | Create project from natural language | devos init "React blog with Tailwind" |
devos run | Run with auto-debugging | devos run --max-retries 5 |
devos goal | Full autopilot mode | devos goal "Build a SaaS app" |
devos auto | Code → Run → Fix loop | devos auto --max-iterations 10 |
devos deploy | Deploy to cloud or local | devos deploy --target=vercel |
devos shell | AI-powered terminal | devos shell |
devos edit | AI code editing | devos edit app.js "Add auth" |
devos test | Auto-generate & run tests | devos test --fix --coverage |
devos refactor | AI code refactoring | devos refactor --perf |
devos security | Vulnerability scan | devos security --fix |
devos watch | Auto-fix on file changes | devos watch |
devos monitor | Self-healing server watch | devos monitor --auto-fix |
devos explain | Explain project or file | devos explain src/auth.js |
devos find | Semantic code search | devos find "login handler" |
devos migrate | Code migration | devos migrate typescript |
devos model | Manage AI models | devos model use gpt-4o |
devos config | Configuration | devos config set KEY=value |
devos env | Environment variables | devos env set DB_URL=... |
devos status | Deployment status | devos status |
devos logs | View logs | devos logs -f |
devos rollback | Rollback deployment | devos rollback v1.0.2 |
devos cost | API cost tracking | devos cost --period=week |
Use cloud APIs, local models via Ollama, or OpenRouter for 25+ models.
GPT-4o, GPT-4o-mini, o1, o3-mini, o4-mini
devos model use gpt-4o
Claude Sonnet 4, Claude Opus 4, Claude 3.5
devos model use claude-sonnet-4-20250514
Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.0
devos model use gemini-2.5-pro
Llama 3.3, CodeLlama, Mistral, Qwen, DeepSeek
devos model use llama3.3
25+ models — GPT, Claude, Gemini, Llama, Mixtral
devos model use openai/gpt-4o
Auto-selects the best model based on task complexity
devos model auto
AI analyzes your project and deploys to the optimal platform automatically.
Detects framework, dependencies, and optimal platform
Generates Dockerfile, build configs, env vars
Compiles, bundles, and optimizes
Ships to cloud and returns live URL
Install DevOS and start building with AI in seconds.
npm install -g devos-ai