← All companies

Agents / Company profile

LangChain

What LangChain does

LangChain is a private company building an “open agent engineering platform” and commercial platform for developing, orchestrating, evaluating, and operating AI agents. On the open-source side, LangChain publishes agent-building frameworks such as LangChain and LangGraph, while on the commercial side it offers LangSmith, a framework-agnostic product for agent observability (tracing), evaluation, and deployment—positioning LangSmith as the place teams iterate on agent quality using live and test data. The company also provides agent “harnesses” such as Deep Agents and corresponding managed offerings intended to reduce the operational burden of running long-horizon, tool-using agent workflows in production. LangChain’s stated focus is enabling companies to “own their intelligence,” emphasizing control over models, context, evaluation, runtime policies, and deployment so that agent systems can be continuously improved rather than treated as one-off prototypes.

Business model: LangChain commercializes LangSmith (and related managed capabilities) as a paid platform. The company offers multiple deployment modes, including cloud and BYOC (Bring Your Own Cloud), to support enterprise security and data-residency needs. Its pricing is consumption- and/or unit-based around LangSmith compute/storage concepts described in its pricing documentation.

Strategic position (as of late August 2026): LangChain is heavily iterating on production-readiness features for agents—adding runtime governance (LLM Gateway), tuned evaluation capabilities (Tuned Evaluators), managed deployment for Deep Agents (Managed Deep Agents), and enterprise-focused blueprints and integrations (e.g., NVIDIA NemoClaw, MongoDB Atlas integration). These efforts suggest the company is competing not only on developer usability of agent frameworks, but also on the end-to-end lifecycle tooling needed to ship reliable agentic systems in enterprise environments.

News

Company record

Aug 18, 2026 · Official · LangChain BlogIntroducing LangSmith Tuned Evaluators, starting with Perceived Error

LangChain introduced Tuned Evaluators in LangSmith, starting with an evaluator for Perceived Error that automatically attaches quality feedback to production traces and threads.

Aug 17, 2026 · Official · LangChain BlogAgentic Commerce at Scale: Your LangChain agents can transact securely

LangChain announced AgentCore Payments integration for LangChain agents, positioning it as middleware that enforces spending limits and supports agent payment workflows.

Aug 12, 2026 · Official · LangChain BlogLangSmith BYOC is now generally available on AWS

LangChain announced general availability of LangSmith Bring Your Own Cloud (BYOC) on AWS, describing a deployment model intended to keep sensitive traces and datasets inside the customer’s AWS environment.

Aug 12, 2026 · Official · LangChain BlogWhy managed agents are the next big thing in agent building

Harrison Chase described the progression to managed agent experiences and how LangChain positions Managed Deep Agents as the developer-facing version of that approach.

Aug 07, 2026 · Official · LangChain BlogManaged Deep Agents is now in Public Beta

LangChain announced Managed Deep Agents as available in public beta, describing a hosted runtime for Deep Agents in LangSmith with managed persistence, memory mounts, sandbox lifecycle, and deployment.

Aug 05, 2026 · Official · LangChain BlogHow we build an autonomous SRE Agent for Kubernetes Deployments

LangChain described building an autonomous SRE agent for Kubernetes, including design choices and how LangSmith is used to make the system trustworthy.

Aug 04, 2026 · Official · LangChain BlogHow to evaluate voice agents: execution, outcomes, and experience

LangChain published guidance recommending evaluation across three dimensions for voice agents and describing how LangSmith tracing and scoring can support an iterative evaluation loop.

Jul 31, 2026 · Official · LangChain BlogEvaluating code review agents with ReviewBench

LangChain described ReviewBench, a code-review evaluation benchmark built from real PR feedback and turned into reproducible Harbor tasks.

Show 4 earlier updates
Jul 30, 2026 · Official · LangChain BlogLangSmith LLM Gateway: runtime governance built into the agent lifecycle

LangChain introduced an LLM Gateway designed to enforce runtime controls such as spend limits and redaction, surfacing policy events as traceable events inside LangSmith.

Jul 08, 2026 · Official · LangChain BlogLangChain and NVIDIA launch the NemoClaw Deep Agents Blueprint

LangChain announced the NemoClaw blueprint for LangChain Deep Agents, intended to enable enterprises to build open and governed agent systems with NVIDIA Nemotron and secure runtime elements.

Apr 08, 2026 · Official · LangChain BlogAnnouncing LangChain + MongoDB Partnership: The AI Agent Stack That Runs On The Database You Already Trust

LangChain announced a partnership with MongoDB, describing integrations across LangSmith and open-source components to support agent state persistence and retrieval workflows tied to MongoDB Atlas.

Mar 16, 2026 · Official · LangChain BlogLangChain Announces Enterprise Agentic AI Platform Built with NVIDIA

LangChain announced an enterprise integration with NVIDIA, describing an agentic AI development platform that combines LangChain frameworks and LangSmith with NVIDIA model and runtime components.

Source map · 2 recurring channels · 13 references

Still resolving: Newsroom · Official X

What it builds

LangChain

An open-source agent development framework intended to help developers build LLM-powered applications and agent systems by composing building blocks.

source ↗
LangGraph

An open-source agent orchestration framework described by LangChain as giving developers low-level control for production agent systems.

source ↗
LangSmith

A commercial, framework-agnostic agent engineering platform for observability (tracing), evaluation, and deployment of LLM applications and agents.

source ↗
LangSmith LLM Gateway

A LangSmith runtime governance layer designed to sit between agents and LLM providers, providing controls such as spend limits, rate limits, and redaction during agent execution.

source ↗
LangSmith Tuned Evaluators

A LangSmith evaluation capability that automatically attaches quality feedback to production traces and threads; LangChain announced Tuned Evaluators starting with an evaluator for “Perceived Error.”

source ↗
LangSmith BYOC (Bring Your Own Cloud)

A deployment model that runs LangSmith capabilities inside a customer’s AWS environment while LangChain operates and provisions the deployment components.

source ↗
Deep Agents

An open-source “agent harness” that LangChain positions as a reusable pattern for long-horizon, tool-using agents (e.g., with durable execution, memory, subagents, and human approval flows).

source ↗
Managed Deep Agents

A managed service that provides a hosted runtime for Deep Agents, intended to reduce the infrastructure and operational work required to run long-running agents in production.

source ↗
NemoClaw for LangChain Deep Agents blueprint

An NVIDIA-developed blueprint intended to help enterprises build open, governed agent systems, combining LangChain Deep Agents components with NVIDIA runtime and a Nemotron model configuration.

source ↗

Key metrics

Milestones & partnerships

Aug 07, 2026
Managed Deep Agents in public beta

LangChain announced Managed Deep Agents is available in public beta for hosting Deep Agents on LangSmith.

Jul 30, 2026
LangSmith LLM Gateway public beta

LangChain announced that LangSmith LLM Gateway was available in public beta, framing it as runtime governance for production agents.

Apr 08, 2026
MongoDB partnership for agent stack integrations

LangChain announced its partnership with MongoDB to connect LangSmith and open-source retrieval/orchestration components to MongoDB Atlas infrastructure.

MongoDB
Mar 16, 2026
Enterprise integration announcement with NVIDIA

LangChain announced an enterprise agentic AI platform built with NVIDIA, positioning it for production-grade agent development.

NVIDIA
Oct 20, 2025
Series B raised for the agent engineering platform

LangChain announced raising $125M at a $1.25B valuation and described expanded platform capabilities spanning LangChain/LangGraph 1.0 and LangSmith workflows.

IVP · Sequoia · Benchmark · Amplify · CapitalG · Sapphire Ventures
Feb 15, 2024
LangSmith general availability and Series A announcement

LangChain announced LangSmith GA and a $25M Series A led by Sequoia Capital.

Sequoia Capital
Jul 01, 2023
LangSmith launched in closed beta

LangChain states LangSmith initially launched in closed beta in July 2023 before GA.

Oct 24, 2022
Initial open-source launch of the LangChain Python package

LangChain describes that the first version of its LangChain Python package was released on October 24, 2022.