EU AI Act · Articles 12 + 14 · high-risk obligations apply 2 December 2027

Prove what your agents decided.

Paraph is the independent, tamper-evident system of record and human-oversight layer for AI agents — across every framework. The record your agents, and the tools that build them, can’t write for themselves.

Watch an agent make a high-risk call — held for a human, recorded, and provable.

EU AI Act · Article 12 + 14 · live decision record

€40,000 consumer loan · applicant #4821

recording
  1. starting the underwriting agent…

Built for the regulator’s questions, not just the developer’s.

Observability tools were built to help engineers debug. A compliance team needs evidence it can hand over and defend — independently verifiable, not vendor-attested — and that an agent framework, by its nature, cannot provide about itself.

01 · Vendor-neutral

One record across every framework

LangGraph, CrewAI, the OpenAI Agents SDK, your own code — one neutral record over all of it. A framework vendor can only ever instrument its own agents.

02 · Independent

The auditor isn’t the audited

Paraph sits outside the stack that makes the decisions — and exported evidence is anchored to an independent RFC 3161 timestamp authority, so the proof doesn’t rest on trusting the agent, the framework, or us.

03 · Audit-grade

A hash chain, not a log file

Every event is cryptographically linked. Alter a recorded decision and the chain breaks — detectably. You just watched it happen above.

04 · For the second line

Made for the people who answer to the regulator

A reviewer queue for live sign-off and a one-click evidence export — for the risk & compliance function, not only the engineers who shipped the agent.

Observability tells you what your agent did. Paraph proves the record of it hasn’t been altered since — and lets a third party check that claim without trusting us. That is the difference between telemetry and evidence.

Mapped to the letter of the Act

Art 12(1)automatic recording of events over the system’s lifetimeevery step appended to the tamper-evident, hash-chained record
Art 14(4)(d)the overseer may decide not to use, disregard, override or reverse the outputthe gate — approve or reject, with the reviewer’s signed note on the record
Art 14(4)(e)the overseer may intervene in the operation or interrupt the systemq.gate() halts the agent until a human resolves — live, in the control room

Article numbering per Regulation (EU) 2024/1689, as amended by the 2026 Digital Omnibus.

“High-risk AI systems shall technically allow for the automatic recording of events (‘logs’) over the lifetime of the system.”

EU AI Act, Article 12(1) · applies to Annex III high-risk systems from 2 December 2027

Dec 2, 2027
Annex III high-risk obligations — credit decisioning and insurance underwriting among them — apply, postponed from August 2026 by the Digital Omnibus (Parliament 16 June, Council 29 June 2026).
€15M / 3%
The fine tier for breaching high-risk obligations such as Articles 12 and 14 — €15M or 3% of worldwide annual turnover, whichever is higher.
No backfill
The deadline moved. The duty didn’t — and records don’t work retroactively. Decisions your agents make now are provable later only if they’re on the record now.

Capture, oversee, prove — in two lines.

Instrument an agent with the SDK (TypeScript or Python), or point your existing OpenTelemetry exporter at Paraph. Nothing about your stack has to change.

Capture

Every step, on a hash chain

Messages, tool calls, model calls, decisions — appended to a tamper-evident record in Postgres.

Oversee

Gate the high-risk actions

q.gate() pauses the agent until a human signs off. Live, in the control room.

Prove

Export the evidence

One click exports a verifiable evidence package — the record, the sign-offs, and an integrity proof anchored outside Paraph — built to support Article 12 and 14 reviews.

const q = paraph({ apiKey, run: "loan-4821", agent: "underwriter" });
q.modelCall("underwriter", "recommend approval, pending sign-off");

// pauses here until a human resolves it
const decision = await q.gate({ action: "Approve €40,000 loan", policy: "block" });

The precise answers.

What does Article 12 of the EU AI Act require?

High-risk AI systems must technically allow the automatic recording of events — logs — over the system's lifetime, so decisions are traceable after the fact. The Act doesn't enumerate what an 'event' is for an agent; Paraph's reading is that a defensible record captures every message, tool call, model call and decision, kept queryable and tamper-evident so it can be examined years later.

When do the high-risk obligations actually apply?

Annex III high-risk obligations — including credit decisioning and insurance underwriting — apply from 2 December 2027, postponed from August 2026 by the Digital Omnibus (adopted by the European Parliament on 16 June 2026 and the Council on 29 June 2026; AI embedded in Annex I regulated products: August 2028). The duty didn't move, and records can't be created retroactively: decisions your agents make now are provable later only if they're on the record now.

How is Paraph different from observability tools like LangSmith or Langfuse?

Observability is telemetry for the engineers who built the agent — made for debugging and optimising, and it does that well. Paraph is made for a different reader: an independent, vendor-neutral record where every event is hash-chained, human sign-offs are first-class, and exports carry an RFC 3161 timestamp from an independent authority — so a third party can verify the record without trusting Paraph, the framework, or the agent.

Does Paraph work with my agent framework?

Yes, by design. Drop-in TypeScript and Python SDKs, a one-line LangChain / LangGraph callback, and OTLP ingestion that accepts any OpenTelemetry exporter with its default settings — so LangGraph, CrewAI, the OpenAI Agents SDK and hand-rolled agents all land on one neutral record.

What makes the record tamper-evident?

Every event's hash covers the previous event's hash and the event's canonical content, so altering, dropping or reordering anything after the fact breaks the chain — detectably, at the exact event. Evidence exports additionally anchor the record's digest with an RFC 3161 timestamp authority, so the proof doesn't rest on trusting our database. You can watch the chain break in the demo above.

Can I check these claims without talking to you?

Yes — that's the point. Download the sample audit pack: a genuine anchored export with offline verifiers you run yourself (node for the hash chain and digest, openssl for the independent timestamp). Or drop any export on the /verify page — verification runs in your browser and the file never leaves your machine. The technical whitepaper walks through the whole construction.

See it on your own agents.

A pilot maps Paraph to one high-risk workflow — instrumented, supervised, and on the record — in weeks.

Book a pilot →

Due diligence first? Sample audit pack ↓ · Whitepaper: Evidence, not telemetry (PDF)