LangWatch

LangWatch is presented as an LLM engineering platform for teams that ship AI to production. Its page describes capabilities for agent simulation testing, LLM evaluation, LLM observability and tracing, an AI Gateway for routing calls through one endpoint with your own keys and budgets, prompt management with versioning and GitHub sync, voice AI testing, LLM red-teaming, and tracking token usage for coding agents such as Claude Code and Codex. It also mentions an AI assistant named Langy that drafts test plans and can turn failures into pull requests. Deployment options listed include cloud (EU, US, UK, APAC), self-hosted via Docker, Kubernetes/Helm, or in a VPC, and hybrid. The page lists ISO 27001 certification, GDPR compliance, RBAC, SCIM and SSO, audit logs, and an open source (Apache 2) mention. Integrations named include OpenTelemetry, LangChain, LangGraph, CrewAI, OpenAI Agents, AWS Bedrock, Azure OpenAI, Vertex AI, DSPy, and Python and TypeScript SDKs.

Key features

• Langy AI engineer that reads traces, writes tests, and opens pull requests
• Instant Evals for searching production history with on-demand evals
• Agentic AI testing with simulated user scenarios in text and voice
• LLM evaluation with LLM-as-a-judge, pairwise comparison, and multimodal scoring
• LLM observability with OpenTelemetry-native tracing, cost and latency monitoring
• AI Gateway routing multiple providers through one endpoint with per-person and per-team budgets
• Prompt management with versioning, A/B testing, and GitHub sync
• Voice AI agent testing and simulation
• LLM red-teaming with adversarial simulations
• Token and cost tracking for Claude Code, Codex, and other coding agents
• Custom trace views, AI search, topic clustering, and analytics graphs
• Self-hosted, hybrid, and cloud deployment options
• RBAC, SCIM and SSO, audit log, and custom retention policies

Best suited for

• AI platform teams
• Teams shipping LLM agents to production
• Engineering teams tracking coding assistant usage and cost
• Organizations needing self-hosted or EU-hosted AI observability
• Product and engineering teams testing voice AI agents

Pricing

Not stated on the website