# LangChain 32 Hacker News commenters recommended LangChain and 79 criticized it, November 2025 to November 2026. It was recommended in 10 of 13 months; owners who said how long report a median of 1 years (3 owners). **Verdict:** Recommended for easy switching between LLM providers and many integrations; critics find it bloated, overly abstract, and built for older models. ## Why people recommend it - Unified API across LLM providers (4 comments): "I could maybe see that being the case for LangChain, specifically, if you're just thinking about abstracting the LLM provider's APIs." ([source](https://news.ycombinator.com/item?id=49727066)) "for all the crap that LangChain gets it at least has its own layer for upstream LLM provider calls" ([source](https://news.ycombinator.com/item?id=47503274)) "I appreciate having a unified API across providers" ([source](https://news.ycombinator.com/item?id=48661127)) - Good starting point for prototypes (2 comments): "The prototype worked great: chain a few calls together, get a response and it felt like we were basically done." ([source](https://news.ycombinator.com/item?id=47426675)) "Langchain is a decent place to start if starting from scratch." ([source](https://news.ycombinator.com/item?id=46257857)) - Built-in prompts and common facilities (2 comments): "they do have good prompts backed into the framework" ([source](https://news.ycombinator.com/item?id=46387297)) "which already has most of the facilities for that" ([source](https://news.ycombinator.com/item?id=46052189)) - Used in production for complex needs (2 comments): "The company I work for uses LangChain heavily, but that's because we have fairly complex requirements" ([source](https://news.ycombinator.com/item?id=47146042)) "It’s deployed and doing what everyone needs it to do." ([source](https://news.ycombinator.com/item?id=46388769)) ## What people warn about - Bloated and adds unnecessary complexity (4 comments): "LangChain is definitely a bloated mess, but it provides structure" ([source](https://news.ycombinator.com/item?id=47073702)) "crazy complex prompt composers (langchain) that don't need to exist any more because the underlying model got so much better" ([source](https://news.ycombinator.com/item?id=47843420)) "most of the functionality actually adds more code to use than simply writing my own direct API LLM calls" ([source](https://news.ycombinator.com/item?id=46387087)) - Designed for older models, outdated (3 comments): "crazy complex prompt composers (langchain) that don't need to exist any more because the underlying model got so much better" ([source](https://news.ycombinator.com/item?id=47843420)) "langchain is stuck in innovator dilemma - it was built for gpt 3.5, or 4, it needs a different design for todays models" ([source](https://news.ycombinator.com/item?id=47099283)) "there used to be a whole complex apparatus around getting models to do chain of thought reasoning, e.g., LangChain" ([source](https://news.ycombinator.com/item?id=46015342)) - Changes too fast, unstable (3 comments): "agents are easy until you build a whole application using an API endpoint that OpenAI or LangChain decides to yank" ([source](https://news.ycombinator.com/item?id=45842814)) "it was evolving so fast that even the latest ai models didn't have even remotely up to date information on it" ([source](https://news.ycombinator.com/item?id=46383258)) "I am afraid it will all go in the LangChain direction with VC funding designs that are not yet ready" ([source](https://news.ycombinator.com/item?id=47892318)) ## Alternatives hackers recommend - [Claude Agent SDK](https://hackerapproved.com/p/claude-agent-sdk.md): 36 recommend - [Mastra](https://hackerapproved.com/p/mastra.md): 32 recommend - [LangGraph](https://hackerapproved.com/p/langgraph.md): 31 recommend - [DSPy](https://hackerapproved.com/p/dspy.md): 29 recommend - [Pydantic AI](https://hackerapproved.com/p/pydantic-ai.md): 27 recommend ## Cite as Hacker Approved, "LangChain", https://hackerapproved.com/p/langchain/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).