LangWatch

LangWatch is an Apache-2.0 platform for running AI in production. It lets teams trace, test, route and govern LLM calls across the company, covering both the agents they build themselves and the coding assistants their engineers use. The project is aimed at AI engineers and platform teams that need visibility and control over model usage.

Its feature areas include LLM operations such as observability, agent testing, evaluations and prompt management, plus tracking of coding agent sessions with cost per pull request and per team and privacy controls. An AI gateway offers a single endpoint compatible with the OpenAI and Anthropic APIs, with virtual keys, budgets and routing, and an AI governance layer sits on top. Repository topics mention DSPy, datasets, simulation testing and low-code evaluation.

LangWatch is written in TypeScript and can be used through LangWatch Cloud or self-hosted; the README says only Node.js is required to try it locally, with separate guidance for production deployments. It suits teams moving LLM prototypes into monitored, tested production systems.