LangSmith
#6 in Observability & EvalsLangchain · 8× · tolest 31. Juli 2026
60
Momentum
LangSmith is a platform for observability and evaluation of LLM and agentic applications, developed by LangChain. It enables tracing, debugging, testing, and monitoring, works framework-agnostically (including with LangChain, LangGraph, OpenAI SDK, Anthropic SDK, LlamaIndex), and supports OpenTelemetry. LangSmith launched as Closed Beta in July 2023 and reached general availability in July 2024; the client SDK is MIT-licensed, while the platform itself is a paid SaaS offering with a self-hosting option for Enterprise customers. The associated deployment component (formerly "LangGraph Platform") was renamed to "LangSmith Deployment" in October 2025.
Momentum-Verloop
16.05.14.08.
Features
| Deployment (Self-Hosted/Cloud) | Managed cloud, bring-your-own-cloud (BYOC), and self-hosted (Kubernetes on AWS/GCP/Azure) available |
| License | Client SDK (langsmith-sdk) under MIT License; platform itself proprietary SaaS |
| Platform | Web app (smith.langchain.com) with Python, TypeScript, Go, and Java SDKs |
| Price | Developer: free (5,000 traces/month); Plus: $39/seat/month (10,000 traces included, overage $2.50/1,000 traces); Enterprise: custom |
| Protocol Compatibility | OpenTelemetry support (ingest and export trace data); native A2A, MCP, and Agent Protocol support in Deployment |
| Release Date | Closed beta July 2023, general availability July 2024 |
| Supported Models/Providers | Framework-agnostic: OpenAI, Anthropic, Vercel AI SDK, LlamaIndex, CrewAI, Pydantic AI, and more, including custom implementations |