LangChain
LangChain is an open-source Python/JS framework for building LLM-powered apps, maintained by LangChain Inc. (founder Harrison Chase) under the MIT license. As of July 2026, the langchain package is at v1.3.13 (July 10) and langchain-core at v1.4.9 (July 8), following the LangChain 1.0 / LangGraph 1.0 stable milestone in October 2025. LangChain Inc. raised a $125M Series B in October 2025 at a $1.25B valuation. PyPI shows 100M+ monthly downloads and GitHub shows 139k+ stars (June 2026) β the most-downloaded agent framework by that measure. Its orchestration layer LangGraph (v1.2.9, 37k+ stars) models agents as durable, cyclic state graphs instead of linear prompt chains. LangSmith tracing is free up to 5k traces/month, then $39/seat/month (Plus tier). Current version/status: langchain 1.3.13 / langchain-core 1.4.9 / langgraph 1.2.9 (July 2026); LangChain 1.0 shipped Oct 2025. License: MIT. Pricing: Core library free/open-source; LangSmith observability free to 5k traces/mo, Plus $39/seat/mo (10k traces), Enterprise custom; LangGraph Platform managed deploys from $39/user/mo with 100k node executions included, $0.001 per extra node execution. Maintained by LangChain Inc. (founder Harrison Chase; $125M Series B Oct 2025, $1.25B valuation).
Pick LangChain/LangGraph when you need the widest model/tool integration catalog and durable, cyclic agent state with human-in-the-loop checkpoints for production. Skip it when retrieval quality, not orchestration, is your bottleneck β LlamaIndex's RAG-first pipelines and LlamaCloud parsing outperform LangChain's bolted-on retrievers. Skip it too if a small team needs a working multi-agent crew in two weeks with no graph-theory learning curve β CrewAI's role/task abstraction ships faster. Even the community's own running joke stands: pre-1.0 LangChain accumulated so much abstraction surface that "framework fatigue" became the top complaint, which is exactly why the October 2025 1.0 rewrite trimmed the API. Known weakness: Pre-1.0 LangChain became notorious in its own community for abstraction bloat and breaking changes ("framework fatigue"), which is why the Oct 2025 1.0 release was a deliberate API-trimming rewrite rather than a routine version bump.
See how it ranks β