Today, LangChain, the startup facilitating the development of large language model (LLM) apps with its open-source framework, announced it has raised $25 million in a series A round, led by Sequoia Capital. The company also said it is launching LangSmith, its first paid LLMOps product for general availability.

Designed as an all-in-one platform, LangSmith allows developers to accelerate their LLM application workflows by covering the entire lifecycle of the project, right from development and testing to deployment and monitoring. It launched in closed beta in July last year and is already being used by thousands of enterprises every month, according to the company.

The move to launch this offering comes as developers need solutions to build applications powered by language models and also advanced visibility and tooling to make sure they are highly performant and reliable in production.

What to expect from LangChain’s LangSmith?

With its open-source framework, LangChain gave developers a much-needed programming toolkit – with a common set of best practices and composable building blocks – to build LLM-powered applications. It can pull in LLMs through APIs, chain them together, and connect them with data sources and tools to accomplish different tasks. The project started as a mere side hustle but evolved quickly to be the backbone of more than 5,000 LLM apps, including internal apps, autonomous agents, games, chat automation and more.

VB Event

The AI Impact Tour – NYC

We’ll be in New York on February 29 in partnership with Microsoft to discuss how to balance risks and rewards of AI applications. Request an invite to the exclusive event below.

 


Request an invite

However, giving the toolkit to build apps is not going to be enough. There are multiple roadblocks in each stage of taking an LLM app to production — which is where LangSmith, the new paid solution, comes in. It enables developers to debug, test and monitor their LLM applications. 

Workflows supported by LangSmith
Workflows supported by LangSmith

When prototyping, developers using LangSmith can get full visibility into the entire sequence of LLM calls and spot the source of errors and performance bottlenecks in real time to debug and iterate. They can collaborate with subject-matter experts to improve the behavior of the app and even layer human feed or AI-assisted evaluation to check for relevance, correctness, harmfulness, insensitivity and more.

Once the prototype is finalized, the unified platform helps users deploy it with hosted LangServe and gives full visibility into what is happening in production, covering everything from costs and latencies to anomalies and errors.

This ultimately enables enterprises to deliver LLM applications that perform well in production, both in terms of quality and cost efficiency.

Significant early adoption

In a blog post announcing the investment, Sonya Huang and Romie Boyd from Sequoia wrote that LangSmith has seen more than 70,000 signups since its closed beta launch in July 2023. Currently, more 5000 companies use the technology every month, including known industry names such as Rakuten, Elastic, Moody’s and Retool.

“Elastic powers their Elastic AI Assistant for security in LangChain and leverages LangSmith for visibility helping them get to production fast. Rakuten relies on LangSmith for rigorous testing and benchmarking so they can make tradeoff decisions about their copilot Rakuten AI for Business, built-in LangChain, in a systematic way. And Moody’s relies on LangSmith for automated eval, easy debugging and experimentation so they can quickly iterate and innovate,” Huang and Boyd noted.

While the technology has already started making waves, its adoption is only expected to soar – now that it is publicly available in a rapidly evolving AI space.

“The team is building in a rich problem space with plenty to explore, and they are guided by a passionate user community with no shortage of important problems for them to solve,” the Sequoia executives added while noting that this is just the beginning for LangChain.

As the next step, LangChain says it plans to introduce multiple capabilities to expand the LangSmith platform. This will include support for regression testing, the ability to run online evaluators on a sample of production data, better filtering and conversation support and easy deployment of applications with hosted LangServe. It will also launch enterprise-grade features to help with administration and security.

With this round from Sequoia, LangChain’s total capital raised has touched $35 million. The previous $10 million round was led by Benchmark, according to Crunchbase data. Other offerings helping with the evaluation and monitoring of LLM apps are TruEra’s TruLens, W&B Prompts and Arize’s Pheonix.

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.

Source link