LangChain

What LangChain is

LangChain is an open-source framework for building applications on top of large language models (LLMs). Its central idea is decomposition: a complicated task is split into smaller components called chains, which can then be connected together ^[4]. Instead of a single prompt-and-response call, a developer describes a sequence of steps — fetch a document, ask a model to work on it, hand the result to another tool — and the framework handles the connections between the pieces. The project is released under the MIT license and lives in the langchain-ai/langchain repository on GitHub, which was created on 17 October 2022 and has roughly 149,000 stars ^[1].

Why it appeared

LangChain began in late 2022 as an open-source project founded by Harrison Chase, who was then an engineer at the machine learning startup Robust Intelligence ^[2]. Business Insider reported that it was launched as an open-source project in October 2022 ^[4]. The gap it addressed was concrete: at the time, LLMs had no access to real-time information and no way to perform actions such as searching the web, calling APIs or interacting with a database ^[2]. A model on its own could produce text, but it could not look anything up or change anything outside the conversation. A framework that let a program call the model, take its output and route it to a search engine, an API or a database addressed exactly that limitation ^[2].

Chains and agents

The chain is the basic unit. A developer writes each step of a task as a component and links the components so that the output of one becomes the input of the next ^[4]. Because the pieces are separate, they can be swapped, reused or arranged differently for a new application, and the model can be asked to choose among them rather than follow a fixed script.

The agent side of the framework grew partly out of work done outside the project. Yohei Nakajima, a general partner, built BabyAGI, a task-driven autonomous agent assembled from GPT-4, Pinecone and LangChain ^[4]. That project prompted Chase to incorporate similar functionality into LangChain while crediting Nakajima ^[4].

LangSmith and the product line

Alongside the open-source framework, the company offers LangSmith, a separate closed-source product for observability, evaluation and monitoring of LLM applications ^[2]. TechCrunch reported that LangSmith had driven annual recurring revenue between $12 million and $16 million ^[2]. By October 2025 the company's announcements covered three pieces of software: the agent builder LangChain, the orchestration and context and memory tool LangGraph, and the testing and observability tool LangSmith ^[3].

Funding

In April 2023 Chase secured a $10 million seed round from Benchmark, followed a week later by a $25 million Series A led by Sequoia, reportedly valuing LangChain at $200 million ^[3]. Business Insider reported the Sequoia round at between $20 million and $25 million and said Sequoia pre-empted the round without a formal fundraising process ^[4]. In October 2025 LangChain raised $125 million at a $1.25 billion valuation in a deal led by IVP, with new investors CapitalG and Sapphire Ventures and existing investors Sequoia, Benchmark and Amplify ^[3]. Around that time, TechCrunch reported that the open-source project had about 118,000 GitHub stars ^[3].

Where it sits

LangChain's position comes from being one of the first widely used ways to put a language model inside a larger program. The chain abstraction gave developers a shared vocabulary for multi-step LLM applications, and the open-source repository gave them a starting point that did not require building the connections themselves. The later product line — a commercial observability and evaluation service, an orchestration layer, and an agent builder — describes a path from a single open-source library, created by one engineer in October 2022, to a company with paid products on top of it ^[1][2][3]. The star counts and the funding rounds are the public record of how many developers adopted the framework and how much investors committed to the company around it ^[1][3].