Introduction
Build agents that can plan, use subagents, and leverage file systems for complex tasks
Build intelligent agents that can plan, delegate, and manage complex tasks.
Built on Vercel AI SDK v6, this library extends basic tool-calling with planning tools, virtual filesystems, subagent spawning, and long-term memory.
Quick Start
🚀 Quickstart
Get up and running in minutes with our quickstart guide
📖 Overview
Learn about Deep Agent, its architecture, and when to use it
💻 CLI Usage
Interactive command-line interface for agent development
Core Features
📋 Planning & Task Decomposition
Built-in write_todos tool enables agents to break down complex tasks and track progress
🗂️ Virtual Filesystem
Stateful file operations across tool calls for context management
🤖 Subagent Spawning
Delegate specialized work to isolated subagents for clean context
🧠 Long-term Memory
Cross-conversation persistence with pluggable backend storage
🛡️ Human-in-the-Loop
Approval workflows for sensitive operations and safety controls
⚙️ Middleware
Composable middleware for logging, caching, telemetry, and custom behavior
Configuration Guides
🗄️ Backend Storage
Configure state management with in-memory, filesystem, or persistent backends
🔧 Customization
Configure prompts, tools, and agent behavior for your use case
📦 Checkpointers
Session persistence and state management patterns
🧠 Agent Memory
Implement RAG and memory systems for your agents
Example Use Cases
- Code Analysis & Refactoring: Analyze entire codebases, write documentation, and refactor code
- Research Assistants: Gather information from multiple sources and synthesize comprehensive reports
- Data Processing: Transform and analyze data with multi-step pipelines
- Content Creation: Plan, research, and generate long-form content
- Automation: Coordinate multiple tools and APIs to complete complex workflows
- Testing & QA: Generate tests, analyze code coverage, and identify edge cases