EN

CONTROLLED PREVIEW · METADATA-ONLY BY DEFAULT

A static workflow-record preview. No live audio analysis, speaker verification, fraud determination or legal judgment.

Voice Agent Provenance / Disclosure & Accountability

Review AI voice workflow provenance, disclosure, and human oversight.

In the target model, Sonolock is designed to structure metadata from an AI voice workflow into an intended reviewable, metadata-only RecordPack: generation, transformation, authorization, disclosure, distribution, review and remediation context. Full integration remains deferred; this controlled preview does not perform live audio analysis, speaker verification, fraud determination, or legal judgment.

It is designed around workflow evidence, not biometric identity.

Explore RecordPack sampleRead record formats

Make the voice workflow traceable in its records.

PROVENANCE

Describe generation and change

Organize provider references and transformation context as metadata. No raw audio is required by default.

AUTHORIZATION

Preserve the attributed context

Keep authorization references, disclosure context and distribution scope visible without claiming consent verification.

ACCOUNTABILITY

Document review and remediation

Attach review notes, gaps and remediation references to the relevant workflow version.

Read the chronology. Inspect the authority references.

RecordPack is a structured workflow record. It brings generation, transformation, authorization, disclosure, distribution, review and remediation context together; it is not merely an audio report or downloadable file.

Attributed authorization

A reference can be recorded without proving that consent or authority was valid.

Disclosure context

Describe where disclosure was expected or referenced; do not invent an observed delivery.

Review scope

A human review note states what was inspected and what remains unknown.

Workflow evidence. No biometric identity.

  • No raw audio by default.
  • No voiceprint.
  • No speaker identification.
  • No biometric identity verification.
  • No automated legal judgment.
  • No fraud verdict.

Metadata lineage and disclosure records do not identify a person, verify consent, decide fraud or make a legal determination.

Seven contexts in one workflow record.

Fictional RecordPack RP–001 · Metadata-only chronology

01 GENERATION

Provider reference listed — externally unverified.

02 TRANSFORMATION

Illustrative edit-context reference attached.

03 AUTHORIZATION

Permission reference requested; validity not established.

04 DISCLOSURE

Intended disclosure context described; delivery unverified.

05 DISTRIBUTION

Illustrative channel scope recorded; no delivery occurs.

06 REVIEW

Sample reviewer note: resolve permission and coverage gaps.

07 REMEDIATION

Remediation context unavailable; no action executed.

Existing review surfaces

Keep each review question in its proper context.

01

Verify

AI Voice Authenticity Signal Assessment

Product preview

A validation-stage framework for examining synthesis, conversion, replay, editing, splicing, and provenance context without identifying a speaker or declaring certainty.

Explore Verify

02

Rights

Voice Rights Incident Readiness

Product preview

A product concept for organizing timelines, authorization context, supporting records, and assessment references before professional review.

Explore Rights

03

Workspace

AI Voice Workflow Readiness

Product preview

A product concept for mapping organizations, roles, providers, governance, human oversight, and Article 50 readiness gaps without certifying compliance.

Explore Workspace

RecordPack preview index

Review the controlled-launch surface.

Next action

Inspect a metadata-only RecordPack.

Explore the fictional sample and record formats. Keep provenance, disclosure and limitations together.

View the RecordPack sample

Known limits

Clarity includes what the system cannot decide.

No speaker identification, biometric recognition, ownership determination, legal determination, fraud determination, or court-evidence claim.

Read all limitations

Future Enforce

A future direction, not an available product.

Narrative only. Any future action layer would require separate evidence, authority, safeguards, and human approval. No active capability or primary action is offered here.