LangChain

LangChain started in October 2022 as an open-source Python and JavaScript library for building applications on large language models. The company now sells an agent engineering stack. The free open-source layer has three parts. LangChain gives quick agent building through a common interface to many model providers. LangGraph is a lower-level runtime for stateful, long-running agents with human-in-the-loop steps. Deep Agents handles complex tasks by planning and spawning sub-agents. The commercial layer is LangSmith. It adds tracing, cost and latency observability, evaluations, Agent Server deployment, sandboxes for code that agents generate, an LLM gateway with spend controls, and Fleet, which lets non-engineers build agents in natural language. Its docs list more than 1,000 integrations covering chat models, vector stores and tools, including OpenAI, Anthropic, Gemini, Bedrock, Pinecone, Chroma and Weaviate. LangSmith is not tied to LangChain code and can trace other stacks. Pricing is seat-based with usage on top. The Developer plan is free for one seat and 5k base traces a month. Plus costs $39 per seat per month with 10k traces. Enterprise is custom and invoiced annually, and adds self-hosted or BYOC deployment, SAML SSO, SCIM, RBAC and ABAC, audit logs and SLAs. Extra traces, deployment compute, sandboxes and Fleet runs are metered in LangSmith Units. The audience is engineering teams rather than business users. The main drawback, raised often in practitioner write-ups and community threads, is the framework's thick abstraction layers. Users say they make debugging and production tuning harder, and that version churn before the October 2025 1.0 release forced costly migrations.