Gentrace Review (2026)

Gentrace has shut down

Short page, because there is not much to evaluate.

Gentrace’s own site states that the company has shut down and the product is no longer available. The code has been released on GitHub under the MIT license.

If you are comparing LLM evaluation platforms, remove it from the list. If you are already on it, you have presumably already found out.

We could not find a specific shutdown date published. If you need one for contractual or audit reasons, your own account records are the only reliable source.

The exit was handled decently

Worth acknowledging, because not every vendor does this.

Releasing under MIT is genuinely generous. It is more permissive than the Elastic License 2.0 that Arize Phoenix uses or the AGPL-3.0 on Langtrace’s server - you can fork it, modify it, run it commercially and relicense derivatives with essentially no obligations.

Compare the alternatives. Literal AI kept only a storage Data Layer and took the dashboards, evals and Playground with it. Humanloop simply switched off. Gentrace handed over the whole thing with no strings.

But released code is not a platform

The distinction matters and it is easy to blur when something is free.

There is no maintainer, no security process and no roadmap. In this category that decays faster than most. Provider APIs change, model identifiers change, SDK interfaces change, and every one of those is an integration that breaks with nobody assigned to fix it.

Adopting the Gentrace source means you now own an evaluation platform. For a team with engineers who can genuinely take that on and a strong reason to prefer this codebase, it is a legitimate choice. For everyone else the maintenance cost will exceed what a living competitor charges within a year.

The idea worth keeping

Gentrace’s framing was sound, and it is worth carrying forward as a requirement rather than losing with the company.

It treated LLM evaluation as part of the test pipeline rather than as a dashboard you visit occasionally. Datasets, test cases and experiment comparison wired into CI, so a quality regression blocks a deploy the way a failing unit test does.

A surprising number of evaluation platforms are analysis tools first and make release gating awkward. When you shortlist a replacement, make CI integration an explicit criterion and ask to see it working.

What to use instead

• Braintrust - the closest commercial equivalent for dataset-driven evaluation and regression testing as a managed product.
• promptfoo - the strongest open-source option for the CI-first workflow specifically; it is built around running evals as tests.
• DeepEval - the other serious open-source choice, and it pairs cleanly with a CI pipeline.

All three are actively maintained with real communities, which is precisely what Gentrace can no longer offer at any price.

A pattern worth noticing

We are not going to speculate about any individual company’s finances. But we now maintain pages for Gentrace, Literal AI and WhyLabs, all shut down; Humanloop, acqui-hired and wound down; and Helicone and Traceloop, absorbed into larger platforms. That is significant turnover inside eighteen months, and it should shape how you buy rather than just what you buy.

The practical defence is portability. Prefer permissive licenses, OpenTelemetry-based tracing and open evaluation formats. If your instrumentation is standard and your eval datasets are yours, a vendor failure costs you a migration. If they are not, it costs you a rewrite.

---

Shutdown confirmed against Gentrace’s own site on 31 July 2026. No specific discontinuation date is published. This is a researched directory entry - we have not instrumented this platform, and given it is shut down, we will not.