← Vetro Attest · Free Art. 12/14 checker · Free EU AI Act tools
Vetro Attest vs. Langfuse
*Engineering comparison, not legal advice. Langfuse features change quickly; check their documentation for the current state.*
Langfuse and Vetro Attest are not competitors in the usual sense. Langfuse helps you build and run LLM applications. Vetro Attest turns a trace you already have into evidence that someone outside your company can check. Many teams will want both.
What Langfuse does well
Langfuse is an open-source LLM engineering platform. Its core is available to self-host, and there is a managed cloud version. Teams use it for:
- Tracing LLM calls, tool calls and multi-step agent runs, with inputs, outputs, latency, token usage and cost;
- Evaluation: scoring outputs with human annotation, LLM-as-a-judge or custom scores, and running datasets against new versions;
- Prompt management: versioning prompts and linking them to the traces they produced;
- Debugging and analytics across sessions and users.
It supports OpenTelemetry-based instrumentation, so traces can flow in from many frameworks. For the day-to-day work of making an agent good, it is a strong choice.
What an observability platform does not give you
Langfuse is built to help *you* understand your system. Evidence for a regulator, an auditor or a customer has a different requirement: it must be convincing to someone who does not trust you.
- The operator controls the data. Whoever runs the Langfuse instance, you in the self-hosted case, can in principle modify or delete stored traces. That is normal for an engineering tool, and it is exactly why a trace exported from it is a claim, not proof.
- No independent timestamp. A trace export shows what the database contained when you exported it. It does not show that it looked the same at the time of the incident.
- No third-party signature. An observability backend is not designed to issue a signed statement, verifiable with a public key, that a given trace was intact at a given time.
- No mapping onto the AI Act. Logging duties (Art. 12 for providers, Art. 19 and 26 for retention) ask what the logs make traceable, not just what they contain.
None of this is a flaw in Langfuse. It is outside what an observability platform is for.
What Vetro Attest adds
Vetro Attest accepts an OTLP/JSON trace with the OpenTelemetry GenAI semantic conventions, as-is, the format many instrumentation libraries emit. It rebuilds the spans into a hash-chained ledger and checks it for free. For 49 USD, one-off, you get:
- a full EU AI Act report (Markdown and HTML) mapping the trace onto logging obligations;
- an ed25519 countersignature by this independent service over the digest, event count and timestamp, which anyone can verify against our published public key without an account.
The signed statement also records that the input was an imported OTel trace rather than a natively signed Vetro ledger, so nobody can pass one off as the other.
Plainly: the countersignature attests that the trace verified at a time. It does not attest that the events are true, and the report is a structured aid, not legal advice or a certification.
When to use both
- Keep using Langfuse (or any observability tool) to debug, evaluate and improve your agent.
- When a run matters, such as a credit decision, a candidate screening or a clinical-adjacent recommendation, export its trace as OTLP/JSON and get it countersigned. Store the attestation next to the trace.
- On an audit or a dispute, hand over the trace plus the attestation. The recipient checks the signature themselves.
Langfuse tells you what your agent did. The countersignature lets someone else believe it.
Try it
- Run the free EU AI Act Art. 12/14 checker to see where your logging stands.
- Paste an OTLP trace on the Vetro Attest home page and verify it free. The full report plus countersignature is 49 USD, no subscription.
- Background: OpenTelemetry GenAI traces for compliance.
See a sample ledger verified in one click, check your own logs free or get the countersigned EU AI Act report.