For risk, audit and legal
Every action, answerable.
Every allow and every refusal leaves a 16-field proof record.
elan1 gives risk, audit and security teams evidence for every action an AI agent takes, not only a policy document. Each action, allowed or refused, leaves a 16-field proof record. At L1 the record is tamper-evident; from L2 it is signed and an auditor can check the evidence bundle offline, without asking elan1.
01Governed by design
The allow sits outside the model. Fail toward a human, never toward an action.
When an agent writes to our books, can you show the write could not have happened without the rule check and the approval — and give us the record?
It depends on the mode, and we say which before you rely on it.
elan1 holds the write credential. No write reaches the system of record without a gate decision and a proof record.
elan1 is reached over MCP or HTTP without the credential. Every write intent is evaluated against your rules and recorded.
We state each integration’s mode in writing before you rely on it.
02What you see
One record per action. Refusals included.
The proof record
Who acted, under which declaration and rule version, at what autonomy level, on which record, when; SHA-256, chained to the one before.
Tamper-evident to you at L1
Recompute the chain and any change to a past record shows. Available
Signed and checkable offline at L2
At L2, a record arrives signed, and your auditor checks it with the free offline verifier, without access to elan1. Available
The Assurance door
Records, refusals and approvals, read without touching the system of record.
- F01tenant_id
- tnt_specimen_01
- F02seq
- 1042
- F03prev_digest
- sha256:9f3c…e41a
- F04write_id
- wr_specimen_7c2
- F05write_type
- journal_entry.post
- F06declaration_id
- decl_month_end_close
- F07declaration_version
- 7
- F08policy_version
- 2027.06.1
- F09actor_id
- usr_specimen_approver
- F10actor_kind
- human
- F11autonomy_level
- A2
- F12record_id
- je_specimen_5531
- F13record_version
- 2
- F14payload_hash
- sha256:51b0…7d9a
- F15lexicon_hash
- sha256:e2c4…0b16
- F16recorded_at
- 2027-07-14T10:42:07+05:30
Illustrative specimen. Not customer data.
03Modes and levels
What is enforced, what is recorded, and at which level.
Switch the level and the mode to see what each promises. L3 Proven is reached when an outside auditor accepts an L2 bundle in writing.
L1 · Guarded · The write was controlled.
No write reaches the system of record without a gate decision and a proof record.
What you can say
“Our agents cannot write to our books without passing rules we set. A person approves anything material. We hold a tamper-evident record of every attempt, including the refusals.”
What you cannot say
- That the record is signed or checkable offline. That comes with L2.
- That someone who does not trust elan1 can check it. At L1, verification needs us.
- That it satisfies a regulation on its own. It serves as an internal control.
Status
Available
Where this is weak Mode A covers only the systems where elan1 holds the credential. Anything written outside them is not covered.
The three levels on elan1 Govern: levels of a claim, not products
elan1 Govern · L1 GuardedThe Right Path, approvals, a single writer and a tamper-evident record you can read.Status: Availableelan1 Govern · L2 AttestedA signature, anchoring, an evidence bundle and a free offline verifier: a signed receipt you can hand to your auditor.Status: Availableelan1 Govern · L3 ProvenA receipt that outlives the vendor.Status: On auditor acceptance04Regulation mapper
Four frameworks, one write path.
Pick a framework, then a row, to see the rung of the Right Path that acts on it, the evidence it leaves, and what is not covered. Not legal advice.
EU AI Act: Record-keeping (Art. 12) maps to rung 8 PROOF.
| Requirement | elan1 control | Evidence produced | Not covered | Enforced at action time? |
|---|---|---|---|---|
| Rung 8 PROOF: a proof record for every governed action, refusals included | Actor F09–F10, rules F06–F08, record F12–F13, time F16 | Logging inside your model or application outside the write | Recorded at the action, both modes | |
| Rung 6 AUTONOMY: named approver, not the requester; autonomy ceiling A0–A5; fail toward a human | Approvals; autonomy F11 | Choosing and training the people you assign to oversight | Mode A enforced; Mode B recorded | |
| Append-only chain; no UPDATE or DELETE on evidence rows | The chain | Your retention schedule, and other law that sets it | After the action | |
| Rungs 3–5: pre-gates G1–G9, versioned rules, guardrail | Declaration version F07 | The risk management system itself | Yes, Mode A only | |
| None | None | Model documentation and instructions for use are yours | No. Outside elan1 | |
| None | None | Not covered. elan1 does not certify you or itself | No. Outside elan1 |
Where this is weak: Mode B records the write; it cannot stop it. “Enforced” applies only in Mode A, where elan1 holds the write credential.
Clause references summarise the official text. Not legal advice. elan1 holds no certificate against any framework. Source: Regulation (EU) 2024/1689, Official Journal of the European Union, 12 July 2024.
Read the full page: EU AI Act: what it asks and how elan1 helps
05The limits, up front
What the record does not prove.
- A check, not a judgementA record shows a check ran and who approved; not that the rule was good.
- Nothing signed at L1Signing is L2 Attested. At L1 the chain is tamper-evident to you.
- No certificatesWe hold no certificates.
- L3 has no dateL3 is reached when an outside auditor accepts an L2 bundle in writing; no date.
- Mode B cannot stop a writeMode B records the write and evaluates it against your rules. It cannot stop it.
06FAQ
Risk, audit and legal ask us.
Evidence that each control ran on each action, not only that a policy exists. elan1 writes a 16-field proof record for every action an agent proposes, allowed or refused: who asked, which rules ran, who approved, what changed and the link to the record before. Refusals sit beside the allows, so an auditor can see the control fire.
Yes, from L2 Attested. Records are signed with the residency domain's Ed25519 key and packed into an evidence bundle that the free offline verifier checks without contacting elan1. At L1 Guarded the chain is tamper-evident to you: recomputing it shows whether anything was changed or removed.
elan1 helps produce evidence for record-keeping and human oversight at the moment of action: EU AI Act Articles 12, 14 and 26, and ISO/IEC 42001 operational control and event logs. It does not classify your system, run your management system or make you compliant. The regulation mapper on this page shows each clause, the rung that acts on it and what is not covered.
We never call the record tamper-proof or immutable, and we say “enforced” only in Mode A, where elan1 holds the write credential. elan1 holds no certificate against any framework, and L3 Proven has no date: it begins only when an outside auditor accepts an L2 bundle in writing. The full list is in the claims register.
At L1 the record is tamper-evident: each record is hashed with SHA-256 and chained to the one before, so recomputing the chain shows any change or removal. From L2 records are also signed and checkable offline.
No. It holds a hash of the payload (F14), not the payload, so personal data stays out of evidence.
No. A model may produce a write intent; it may never produce an allow. evaluate() applies versioned rules with no model, and above the autonomy ceiling a named person who is not the requester decides.
We state each integration's mode in writing before you rely on it. Only Mode A is described as enforced.
Yes. Ask through the contact page with the security review topic, and a person replies within 2 working days with the security review pack.
// FOR RISK, AUDIT AND LEGAL
Watch a refusal write its own record.
The walkthrough shows the check, the approval and the record, one rung at a time. You advance it.