LangWatch
Amsterdam-built, Apache-2.0 platform for tracing, evaluating and simulating LLM agents, in EU cloud or self-hosted
Quick Overview
| Company | Reasoning Engines B.V. |
|---|---|
| Category | LLM Observability and Evaluation Tools |
| Headquarters | Amsterdam, Netherlands |
| Founded | Not stated |
| EU Presence | Yes - Netherlands (EU) |
| Data Location | Cloud regions EU, US, UK and APAC; or your own infrastructure |
| Open Source | Apache-2.0 |
| GDPR Compliant | Yes |
| Self-Hosting | Yes - Docker, Kubernetes/Helm or your own VPC |
| Main Features | LLM tracing, evaluations, agent simulations, prompt management, AI gateway, red-teaming, voice agent testing |
| Pricing | Free Developer plan (50k events/month); Growth EUR 29 per core seat per month; Enterprise on request |
| Best For | Teams that want tracing, evals and simulated users in one open-source tool |
| Replaces | LangSmith, Arize, Helicone |
Detailed Review
LangWatch covers the loop that most agent teams build by hand: trace what the agent did, score it with evaluators, and replay realistic user scenarios before shipping a change. The code is published under Apache-2.0 on the company's GitHub organisation, so the same product can be run as a managed service or inside your own cluster.
The contracting party named in the terms is Reasoning Engines B.V., registered in Amsterdam, and the terms are governed by the Dutch Civil Code. The cloud can be placed in an EU region, which keeps trace data inside Europe for teams that handle customer conversations.
Tracing and evaluation together
Traces show each agent step with cost and latency, and evaluators score the output on the same data. The pricing page lists unlimited evals, prompts and simulations on the Growth plan, so quality checks are not metered separately.
Agent simulations use an LLM to play the user over several turns, with a judge that scores each conversation against criteria you write in plain language. The tests run in pytest or vitest, which makes them usable in CI.
Deployment choices
The vendor lists a managed multi-tenant cloud in EU, US, UK and APAC regions, self-hosting with Docker or Kubernetes/Helm, and a hybrid mode where the data plane runs on your infrastructure and the control plane on theirs. Self-hosting is described as taking about 15 minutes on the homepage.
Pricing model
Developer is free with 50k events per month and 14-day data access for two users. Growth costs EUR 29 per core seat per month with 200k events included, then EUR 5 per 100k events, and 30-day retention with extra storage at EUR 3 per GB.
Limitations to Consider
Pricing combines seats and events, so a high-volume agent with a small team can cost more than a flat plan. The free tier keeps data for 14 days and Growth for 30 days unless you pay for storage. The AI gateway and Langy features are marked new on the homepage. If you mainly need pre-release red teaming, Giskard is more specialised.
Who Should Use LangWatch
Engineering teams that want an open-source LLM observability and evaluation stack with a Dutch contracting party and the option to run it themselves.
Pros and Cons
Pros
- Apache-2.0 code published by the vendor
- Tracing, evals and simulations in one product
- EU cloud region or self-hosting
- Free plan with 50k events a month
- Evals and simulations unlimited on paid plans
Cons
- Seat-plus-event pricing needs estimating
- Short retention on free and Growth plans
- Gateway and Langy features are new
- Enterprise terms and SLAs require sales contact
Alternatives to LangWatch
Other European llm observability and evaluation tools worth comparing before you decide:
Latitude
Spain · Open-source agent observability that groups failures into issues
Galtea
Spain · Test generation, simulation and evaluation for regulated industries
Giskard
France · Red teaming and evaluation for LLM agents before release
Orq.ai
Netherlands · AI gateway with built-in traces, evaluation and governance
See all European llm observability and evaluation tools or compare them directly against LangSmith on our alternatives page.
Frequently Asked Questions
Yes. The vendor's GitHub repository lists the Apache-2.0 licence, and the pricing page offers a self-managed option alongside the cloud.
The homepage lists EU, US, UK and APAC cloud regions, and you can also self-host with Docker, Kubernetes or in your own VPC.
Developer is free. Growth is EUR 29 per core seat per month plus EUR 5 per 100k events after the first 200k. Enterprise is priced on request.