The LangChain team is preparing to release version 1.0 of its core langchain Python package. Eugene Yurtsev, a LangChain maintainer, posted the announcement in the official LangChain Forum Announcements category, stating: “We’re getting ready to release version 1.0 of the langchain package, and we’d love your input before it goes live.” The thread has been active with community feedback, with the latest replies posted within the last 16 hours.
The Restructure
The 1.0 release reorganizes how developers interact with the package. According to the forum post, the changes include:
Re-exporting core primitives like messages, tools, and prompts from langchain_core so developers no longer need to learn multiple package import paths upfront. Common building blocks, including agents, chains, MCP (Model Context Protocol), and retrievers, move to top-level exports for easier discovery.
The package will add langgraph as a formal dependency alongside langchain-core. LangGraph handles agent workflow support, and making it a first-class dependency signals that agentic workflows are now central to the framework’s architecture rather than an optional add-on.
Deprecated modules are being removed. Documentation is being consolidated across the langchain, langgraph, and related packages to reduce duplication and unify coverage of end-to-end workflows that span multiple components.
Deep Agents: The Reference Agent Harness
Alongside the v1.0 preparation, LangChain published Deep Agents, a new open-source agent harness built on LangChain and LangGraph. The MIT-licensed repository, published April 15, provides an opinionated, ready-to-run agent with planning tools, filesystem access, shell execution with sandboxing, sub-agent delegation, and automatic context management including conversation summarization when conversations get long.
Deep Agents also ships a CLI that provides a terminal-based coding agent comparable to Claude Code or Cursor, powered by any LLM. The harness supports MCP via langchain-mcp-adapters and works with any model that supports tool calling, including both frontier and open models.
The timing suggests Deep Agents is part of the v1.0 ecosystem expansion: a canonical reference implementation demonstrating how agents should be built on the restructured LangChain 1.0 package.
What Developers Should Expect
The v1.0 designation for a framework as widely adopted as LangChain carries specific implications. API stability guarantees will follow, meaning breaking changes will be managed with proper deprecation cycles. Enterprise teams that require stable interfaces before committing to a framework dependency will have a versioned contract to build against.
The package restructure will likely require developers to review their import paths. Modules currently imported from langchain_core will be accessible directly from langchain, while deprecated modules will need to be migrated before the release.
The LangChain team is actively soliciting feedback before the release goes live. Developers building on LangChain should review the forum thread and flag any concerns about the proposed restructure while the API is still open to changes.