LangChain vs LangSmith

LangChain

LangSmith

LangChain · 18 features

• Agents — Tool-using agents with reasoning loops.
• Chains (LCEL) — LangChain Expression Language — pipe primitives into chains.
• Checkpointers (LangGraph) — Persist agent state to SQL, Mongo, Redis, Postgres.
• Document Loaders — 150+ loaders for PDF, HTML, Notion, Google Drive, S3, GitHub, etc.
• Human-in-the-loop — Pause agent for approval, then resume.
• LangGraph — Stateful graph-based agent runtime. Durable, replayable, human-in-the-loop.
• LangGraph Platform — Managed hosting for LangGraph agents with state persistence.
• LangGraph Studio — Desktop IDE for debugging agent graphs.
• LangServe — Deploy chains as FastAPI endpoints.
• Memory — Buffer, summary, entity, vector memory stores.
• Output Parsers — Structured JSON, Pydantic schemas, function calling.
• Prompt Templates — Templating + partial filling + output parsers.
• RAG (Retrieval-Augmented Generation) — Standard patterns + 50+ retrievers.
• Streaming — First-class streaming at every layer.
• Subgraphs — Compose agent graphs hierarchically.
• Text Splitters — Recursive, token, semantic splitters for chunking.
• Tools — 400+ pre-built tools (web search, code, databases, APIs).
• Vector Store Integrations — 60+ vector DBs (Pinecone, Chroma, Weaviate, PGVector, Qdrant, Milvus).

LangSmith · 14 features

• Alerts — Threshold alerts on latency, cost, eval metrics.
• Annotation Queues — Human-review workflows for trace quality rating.
• Custom Dashboards — Aggregate metrics dashboards per project/tag.
• Datasets — Collect examples → use as eval sets or training data.
• Evaluations — LLM-as-judge, embedding similarity, custom Python evaluators, offline batch eval…
• LangChain Integration — Auto-trace any LangChain/LangGraph run with env var.
• LangGraph Integration — First-class trace + eval for LangGraph agents.
• LLM Tracing — Automatic trace every LLM call + tool call + chain step.
• OpenTelemetry Export — Export traces as OTLP to Datadog/Honeycomb/etc.
• Playground — Test prompts + models inline before deploying.
• Prompt Canvas — Visual prompt editor with live test + eval.
• Prompt Hub — Public + private prompt library with versioning.
• Self-Hosted (Enterprise) — Docker + k8s deployment in your infra.
• Threads + Sessions — Group traces into conversational sessions.