Langchain

LangChain’s suite of products supports AI development

4.9•116 reviews•

1.1K followers

LangChain’s suite of products supports AI development

4.9•116 reviews•

1.1K followers

Visit website

LLMs

•

Unified API

•

AI Infrastructure Tools

•

AI Metrics and Evaluation

•

LLM Developer Tools

Langchain Launches

Langchain AI Email AssistantThe prototype of ambient agents

Launched on January 20th, 2025

LangSmith General AvailabilityLLM application development, monitoring, and testing

Launched on February 15th, 2024

LangChain TemplatesA collection of production-ready reference architectures

Launched on October 31st, 2023

LangSmithBuild and deploy LLM applications with confidence

Launched on July 20th, 2023

Forum Threads

p/langchain

•

12mo ago

LangChain announces Series B raising $125M at a $1.25B valuation

From @hwchase17 :

Today, we re announcing we ve raised $125M at a $1.25B valuation to build the platform for agent engineering. We re also releasing new capabilities to accelerate the path to reliable agents, including LangChain and LangGraph 1.0 releases, a new Insights Agent, and a no code agent builder. IVP led the round alongside existing investors Sequoia, Benchmark, and Amplify, as well as new investors CapitalG and Sapphire Ventures.

Imagine that they got their start here when I hunted @LangSmith in 2023.
Man, shoulda invested back then!

View all

4.9

Based on 116 reviews

Review Langchain?

Reviewers see LangChain as a broad, useful framework for building LLM apps, especially when teams need many integrations, model switching, RAG, tools, and multi-step agent flows without wiring everything by hand. Several also say productivity rises once the concepts click and community support is strong. The main criticism is complexity: abstractions can feel heavy, debugging is harder, docs can lag, and upgrades may break things. Some users also dislike growing reliance on LangGraph Platform, while founders of tools like Databox praise LangGraph’s control over stateful workflows.

+113

Summarized with AI

Pros

agentic workflow support (20)

LangGraph framework (11)

model integration (11)

LangSmith observability (7)

RAG workflows (6)

rapid prototyping (6)

scalable AI development (6)

flexible framework (5)

complex data source integration (4)

community support (3)

Pros

Cons

Reviews

All Reviews

Most Informative

•3 reviews

What's great

flexible framework (5)rapid prototyping (6)agentic workflow support (20)model integration (11)complex data source integration (4)RAG workflows (6)

Model integration is really the killer feature for us swapping between providers (we've gone between Anthropic and OpenAI more than once depending on the task) without rewriting the whole pipeline saved real time. RAG workflows and the retriever abstractions are solid enough that we didn't have to hand-roll our own chunking and embedding logic in the early months, which mattered when we were still validating the product idea instead of infra. Community support is underrated too when something breaks, there's almost always a GitHub issue or forum thread that's already been through it.

What needs improvement

The abstraction layers can get in the way once you actually need to know what's happening under the hood debugging a chain five layers deep is harder than debugging a raw API call, by design. And LangGraph Platform, the hosted piece, has gotten more gated than the open framework itself, so it's not always clear where the free tier ends and where you're expected to pay.

vs Alternatives

We looked hard at LlamaIndex first, since it's more purpose-built for retrieval, but LangChain's broader agent and tool-calling surface won out once we realized we needed more than a RAG pipeline we needed the model taking actions, not just answering from documents. We also seriously considered skipping a framework entirely and building directly against the Anthropic and OpenAI SDKs, which is tempting early on because it feels like less overhead, but that broke down the moment we needed to swap models mid-project or add a second agent step. LangChain's abstractions cost you some directness, but they bought us the ability to change our minds about the architecture without a rewrite, which mattered more at our stage than raw performance would have.

Ratings

Ease of use

Reliability

Value for money

Customization

Report

92 views3mo ago

•1 review

What's great

LangChain is fantastic in the sense that it has a huge ecosystem, many integrations, and it helped move the whole LLM app space forward. I respect that part.

What needs improvement

For me, the biggest issue is developer experience.

LangChain often feels too deep, too wrapped, and over-designed. Simple things can require too much code and too much framework-specific knowledge.

I would love to see a much simpler default path, clearer docs, less boilerplate, and more transparency around the actual agent loop.

vs Alternatives

I also considered using simpler agent SDKs, like openai-agents-sdk\claude agent sdk.... and direct model APIs. For many common agent use cases, those feel much lighter and easier to control.

Ratings

Ease of use

Reliability

Value for money

Customization

Report

107 views5mo ago

Exstats

•7 reviews

What's great

LangSmith observability (7)LangGraph framework (11)agentic workflow support (20)

I use LangChain and their implementation of DeepAgents in my products, and I can honestly say this is a huge contribution to the AI community!

It makes building capable agents much more practical. Memory, skills, reasoning, and todos well designed and usable in actual products. I know it was inspired by Claude, but the implementation is excellent and under open-source. Well done.

Debugging with LangSmith is a special kind of love haha. Love it! Highly recommend it.

What needs improvement

Do not automatically include all LANGSMITH_* and LANGCHAIN_* env variables in LangSmith traces when tracking is enabled.

Ratings

Ease of use

Reliability

Value for money

Customization

Report

167 views7mo ago

•3 reviews

What's great

flexible framework (5)rapid prototyping (6)agentic workflow support (20)model integration (11)complex data source integration (4)RAG workflows (6)

What needs improvement

vs Alternatives

Ratings

Ease of use

Reliability

Value for money

Customization

Report

92 views3mo ago

•1 review

What's great

LangChain is fantastic in the sense that it has a huge ecosystem, many integrations, and it helped move the whole LLM app space forward. I respect that part.

What needs improvement

vs Alternatives

I also considered using simpler agent SDKs, like openai-agents-sdk\claude agent sdk.... and direct model APIs. For many common agent use cases, those feel much lighter and easier to control.

Ratings

Ease of use

Reliability

Value for money

Customization

Report

107 views5mo ago

Exstats

•7 reviews

What's great

LangSmith observability (7)LangGraph framework (11)agentic workflow support (20)

I use LangChain and their implementation of DeepAgents in my products, and I can honestly say this is a huge contribution to the AI community!

What needs improvement

Do not automatically include all LANGSMITH_* and LANGCHAIN_* env variables in LangSmith traces when tracking is enabled.

Ratings

Ease of use

Reliability

Value for money

Customization

Report

167 views7mo ago