LangChain: The platform for the enterprise AI agent era

As agents become the standard interface for accessing AI capabilities in the enterprise, software development to build these agents is changing. Every user input and agent configuration can produce unpredictable results; new tools are needed to solve these new challenges.

For developers, the answer to improving outcomes is an infrastructure and tooling upgrade, a new, end-to-end agent-specific software development lifecycle that combines product, engineering and data science thinking: agent engineering. LangChain has emerged as the leading platform for building, evaluating, observing and deploying AI agents. That’s why we're leading the company's $125 million Series B.

Grounded founders with great timing

We've gotten to know LangChain co-founders Harrison Chase and Ankush Gola over the past two years and are impressed with their talent, humility and market timing. Harrison invented the agent framework category in October 2022 by solving problems he experienced firsthand. When developers needed more control, LangChain launched LangGraph, a framework for complex agent orchestration. When they needed observability, the company shipped LangSmith, a platform for testing and monitoring agents. As production workloads scaled and enterprises required a purpose-built infrastructure for agents, Harrison and Ankush evolved LangSmith to include deployment, expanding it to become the comprehensive platform for agent engineering.

They were just getting started.

Neutral, open source and end-to-end

LangChain is unique in creating an open source, model-agnostic, end-to-end platform that developers love and trust. One of the platform’s winning characteristics is neutrality. It’s the same concept that led us to back Datadog’s cloud-neutral observability and CrowdStrike’s platform-neutral security, so CTOs aren’t locked into a single tech platform or stack. LangChain supports more than 70 model providers.

With 110,000+ GitHub stars, 4,000+ contributors, and 1,000+ integrations, LangChain has built network effects that compound with every new model provider and every enterprise customer. The open source frameworks — LangChain for pre-built agent architectures and LangGraph for production-grade orchestration — have become the first- and second-most popular agent frameworks in the market, respectively.

Enterprise customers want LangChain

In LangChain, we see a category leader and platform that already has 90 million combined monthly downloads. Even more impressive for an open source software company is its commercial strength in the enterprise: LangChain counts a third of the Fortune 500 as active users. Klarna uses LangChain to build customer support agents. Morningstar uses it to create investment-research agents. Vanta, Rippling, Cloudflare and Workday are each building production agents with LangChain's products.

Those production agents need infrastructure that works across models, scales with complexity and provides enterprise-grade reliability. LangChain has the community, the customers, the product suite and the momentum. And it has Harrison, Ankush and a strong growing team — in other words, everything LangChain needs to become the category-defining AI infrastructure platform.