MIT-licensed database and workspace where AI agents store data, docs, skills and apps, available on desktop, self-hosted or in Busabase Cloud.
Open source
LangChain Deep Agents
Open-source LangChain agent harness: long-term memory, a virtual filesystem, subagents and planning included
What is LangChain Deep Agents?
LangChain Deep Agents is an open-source agent harness for Python and TypeScript. Instead of wiring every component yourself, you call a single function such as create_deep_agent with a model, a list of tools and a system prompt, and get an agent that can plan, read and write files, run code in a sandbox, and delegate work to subagents that run in isolated context windows. The harness manages growing context by summarizing history and offloading large results, supports human approval before sensitive actions, and can load memory and skills at the right moment. Tools can be custom functions, LangChain tools or any server using the Model Context Protocol. It works with models from several providers, and the docs say it is used in production by OpenSWE and LangSmith Fleet. The library is free to use under the MIT license, while managed deployment of deep agents is a separate LangSmith offering.
Use cases
- Build long-running multi-step agents
- Delegate subtasks to isolated subagents
- Give agents a virtual filesystem for context
- Add human approval to risky agent actions
- Connect agents to MCP servers and tools
- Run a terminal coding agent with dcode
Who it's for
Why use LangChain Deep Agents
- Planning, subagents, files and memory come bundled, so you do not assemble them one by one.
- Subagents keep large intermediate work out of the main agent's context window.
- Approval pauses let a person review sensitive tool calls before they run.
- The same stack as LangChain and LangGraph means existing tools and integrations carry over.
- The model is swappable across providers by changing one parameter.
LangChain Deep Agents features
- create_deep_agent entry point for Python and TypeScript
- Task planning and subagent spawning in isolated context windows
- Virtual filesystem with pluggable backends
- Filesystem permissions controlling readable and writable paths
- Sandboxed shell execution and an in-process JavaScript interpreter
- Context management through history summarization and result offloading
- Long-term memory and skills loaded when needed
- Human-in-the-loop approvals at decision points
- Model Context Protocol support for external tools
- Typed event streams for messages, tools and delegated tasks
LangChain Deep Agents integrations
How to get started with LangChain Deep Agents
- Install the Deep Agents package for Python or TypeScript.
- Call create_deep_agent with a model, your tools and a system prompt.
- Invoke the agent with a user message; it plans, uses files and spawns subagents as needed.
- Add approvals, memory, skills or sandboxes, and trace runs in LangSmith if you want observability.
LangChain Deep Agents limitations
- It is a developer library, so using it requires writing code
- You supply your own model access and API keys
- Managed hosting of deep agents is a LangSmith product with its own pricing
- Optional capabilities like task planning and skills extend the base harness
LangChain Deep Agents pricing
LangChain Deep Agents is open source, so you can use and self-host it for free.
Prices can change and may vary by country or exclude tax. Check the official pricing page.
LangChain Deep Agents alternatives
The main LangChain Deep Agents alternatives listed on Debuly are Claude Agent SDK, OpenAI Agents SDK, CrewAI, AutoGen, LlamaIndex and LangGraph. The table compares pricing and platforms; order is A–Z.
| Tool | Pricing | Starting price | Platforms |
|---|---|---|---|
| LangChain Deep Agents | Open source | — | Python, TypeScript, CLI |
| AutoGen | Open source | — | API, CLI |
| Claude Agent SDK | Pricing unknown | — | Python, CLI, macOS, Linux |
| CrewAI | Open source | — | Web, API, CLI, Self-hosted |
| LangGraph | Open source | — | Self-hosted, Web, API |
| LlamaIndex | Freemium | $50/mo | Web, API, Self-hosted |
| OpenAI Agents SDK | Open source | — | Python, macOS, Linux, Windows |
AutoGen
Microsoft's MIT-licensed Python and .NET framework for multi-agent AI apps, now archived; new projects are pointed to Microsoft Agent Framework.
Open sourceVisit AutoGen (official site)
Claude Agent SDK
Anthropic's Python SDK for building agents on the Claude Code runtime, with built-in tools, custom tools, hooks and MCP servers.
CrewAI
Multi-agent framework and platform for building AI agent crews in Python or a visual editor, then deploying and monitoring them.
Open sourceVisit CrewAI (official site)
LangGraph
Low-level open-source orchestration framework and runtime for building stateful, long-running AI agents as graphs.
Open sourceVisit LangGraph (official site)
LlamaIndex
Open-source framework plus the LlamaParse platform for parsing, extracting and indexing documents so LLM apps can use them.
FreemiumVisit LlamaIndex (official site)
OpenAI Agents SDK
Open-source Python framework for building multi-agent workflows with tools, handoffs, guardrails, tracing, sandboxes and voice agents.
Open sourceVisit OpenAI Agents SDK (official site)
LangChain Deep Agents FAQ
Is Deep Agents free and open source?
Yes. The library is open source and MIT licensed, so there is no price to use it. You pay your model provider, and managed deployment through LangSmith is priced separately.
How much does LangChain Deep Agents cost?
The framework itself costs nothing. Costs come from the language model you call and, optionally, LangSmith services for tracing or managed hosting.
What languages does Deep Agents support?
The docs provide Python and TypeScript versions, and there is a terminal coding agent called dcode built on the Deep Agents SDK.
How do Deep Agents differ from LangChain and LangGraph?
LangChain provides core building blocks, LangGraph handles low-level orchestration, and Deep Agents is a complete harness on top with planning, subagents and memory built in.
Does Deep Agents support MCP?
Yes. You can pass tools from any Model Context Protocol server through the tools parameter, alongside custom functions and LangChain tools.
What are alternatives to Deep Agents?
Common alternatives include the Claude Agent SDK, OpenAI Agents SDK, CrewAI, AutoGen and LlamaIndex agents.
Sources & verification
Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.
Image credits
- Logo: from docs.langchain.com · fetched · Vendor trademark, used only to identify the product; removed on request
- Cover: none collected yet, so a generated tile is shown.
Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.
More in Developer platforms 4
All developer platforms toolsOpen-source runtime that limits what AI agents can do, using security policy compiled from a plain-English constitution.
Open sourceChange the g in a GitHub address to a u to view a whole repository as one block of text you can paste into an LLM.
Vercel is a cloud platform for deploying web apps and AI applications, with a free Hobby plan, a Pro plan and usage-based pricing.
Freemium
