LangWatch
OpenTelemetry-native LLMOps platform for tracing, evals and agents
Overview
LangWatch is an Amsterdam-built, open-source LLMOps platform combining real-time OpenTelemetry-native tracing, offline/online evaluations, DSPy-based prompt optimization and agent simulation testing, deployable as free cloud, self-hosted or on-prem.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Tracing, evals and simulations
- Community support via GitHub/Discord
- Unlimited lite users
- Unlimited simulations, evals and prompts
- 30-day retention
- Private Slack/Teams support
- Custom SSO/RBAC and audit logs
- ISO 27001 reports and InfoSec reviews
- Forward-deployed engineer
- AWS/Google Marketplace billing
What you can produce with LangWatch
- Real-time OpenTelemetry-native tracing
- Offline and online LLM evaluations
- DSPy-based automated prompt optimization
- Agent simulation testing before deployment
- Dataset curation and human-in-the-loop annotation
- Built-in AI gateway with virtual keys, budgets and fallback
- Free, self-hosted, or on-prem deployment options
- ISO 27001 reporting for Enterprise customers
ASEAN Perspective
LangWatch in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
LangWatch's most interesting move isn't a feature, it's the pricing shift it made in February 2026: away from per-trace billing (the model Langfuse and LangSmith both use, which can spike hard as usage grows) toward seat-plus-event pricing. LangWatch claims a team running 100k traces a month lands around $80-150, versus $400+ under typical trace-based pricing elsewhere, and stays in the $200-300 range even at 1M traces monthly.
Feature-wise it covers the same ground as Langfuse and LangSmith — tracing, evaluation, dataset curation, human annotation — but adds agent simulation testing and DSPy-driven prompt optimization, plus its own AI gateway with virtual keys and budgets bundled in, more scope than most single-purpose observability tools attempt.
Founded by Rogerio Chaves and Manouk Draisma out of an Antler residency in Amsterdam, backed by a €1M pre-seed (Passion Capital, Volta Ventures, Antler). Good fit for teams building complex, multi-step agents who want to simulate and evaluate before shipping; teams already deep in the LangChain ecosystem may still find LangSmith's native integration hard to beat.
What people say
LangWatch sits in a three-way comparison that shows up constantly in 2026 LLMOps coverage: itself, Langfuse and LangSmith. The dividing lines are consistent across sources — Langfuse is MIT-licensed and treats self-hosting as a first-class deployment option; LangSmith is proprietary and tightly coupled to the LangChain/LangGraph ecosystem, with self-hosting gated behind an Enterprise license; LangWatch positions itself as open-source and framework-agnostic like Langfuse, but adds agent simulation testing and prompt optimization (via DSPy) that neither competitor bundles natively.
The GitHub repo (langwatch/langwatch) has around 3,300 stars — meaningfully behind Langfuse's much larger open-source following, but an active project with regular releases rather than an abandoned experiment.
The pricing restructuring in February 2026 is the most concrete differentiator: LangWatch moved off trace-based billing, framing it explicitly as a reaction to how expensive trace-volume pricing gets for teams that scale usage — a real pain point across the LLM observability space as agent workloads generate far more trace volume than simple chat completions did. Independent cost comparisons back the general shape of the claim: a mid-size team at 100k traces/month landing in the low hundreds of dollars, versus several hundred more under a pure per-trace model.
The Developer tier is free forever with 50k events/month and 14-day retention — enough to evaluate seriously before paying. Growth tier is €29 per core seat plus €5/100k events beyond the included 200k, with 30-day retention. Enterprise adds hybrid/self-hosted/on-prem deployment, custom SSO/RBAC, audit logs, ISO 27001 reporting and a forward-deployed engineer, priced on request.
Independent review-site coverage (G2/Capterra star ratings) is less visible than for the bigger incumbents — most of what's public comes from LangWatch's own comparison pages and third-party "best LLMOps tools" roundups rather than large volumes of verified user reviews, so treat the feature comparisons as directionally accurate rather than independently audited. For a team choosing between the three, the practical filter is usually deployment model and existing stack: LangChain shops lean LangSmith, self-hosting purists lean Langfuse, and teams that want agent simulation plus a bundled gateway without adding a fourth vendor lean LangWatch.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including LangWatch's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with LangWatch unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to LangWatch directly →
Spotted something out of date? Suggest an update →
LangWatch in the news
Developer Tools
Miasma: How a Hijacked CI Pipeline Shipped Malicious npm Packages With Valid Provenance
Developer Tools
OpenTelemetry Graduates CNCF: 12,000 Contributors Lock In the Industry's Unified Observabi...
Developer Tools
GitLab 19.0 drops Redis for Valkey by default and brings Gitaly to Kubernetes
More in Code & Dev Tools
