deepagents logo

deepagents

Free

Agent harness built with LangChain and LangGraph. Equipped with a planning tool, a filesystem backend, and the ability to spawn subagents - well-equipped to handle complex agentic tasks.

Model APIsFreeFree tier
Inputs: textOutputs: text
Type
Open Source

About deepagents

Deep Agents is an open-source agent harness built with LangChain and LangGraph that provides built-in capabilities for task planning, filesystem context management, subagent spawning, and long-term memory. It enables users to build LLM-powered agents that can take actions via tools, read/write files, execute code, connect to external data (memories, skills, domain knowledge), manage growing context through summarization, parallelize tasks by delegating to subagents, pause for human approval, and improve over time by updating memory and prompts. It supports multiple model providers including Google, OpenAI, Anthropic, OpenRouter, Fireworks, Baseten, and Ollama.

Key Features

Task planning
Filesystem backend for context management
Subagent spawning for parallelization
Long-term memory
Tool usage (read/write files, execute code)
Human-in-the-loop approval
Model support: Google, OpenAI, Anthropic, OpenRouter, Fireworks, Baseten, Ollama
Built with LangChain and LangGraph
Context management via summarization and offloading
Updatable memory, skills, and prompts

Pros & Cons

Pros
  • Easiest way to build agents according to documentation
  • Built-in planning and context management
  • Supports multiple LLM providers
  • Open source and free
  • Subagent spawning for parallel tasks
  • Human-in-the-loop capabilities

Best For

Complex multi-step tasksBuilding LLM-powered applications

Alternatives to deepagents