> ## Documentation Index
> Fetch the complete documentation index at: https://docs.meshai.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# EU AI Act Coverage

> Article-by-article mapping of EU AI Act requirements to MeshAI features

The EU AI Act (Regulation 2024/1689) is phasing in on a staggered timeline. Prohibited practices have applied since **February 2025** and GPAI obligations since **August 2025** - both already in force. Under the Digital Omnibus (adopted by the Council on 29 June 2026, pending publication in the Official Journal), the transparency-obligation grace period runs until **2 December 2026**, Annex III high-risk obligations now apply from **2 December 2027**, and Annex I high-risk obligations from **August 2028**. MeshAI maps its features to specific articles so you can demonstrate compliance ahead of those dates.

<Note>
  Dates above reflect the Digital Omnibus as adopted by the Council on 29 June 2026; they are pending publication in the Official Journal and could still shift before then.
</Note>

## Coverage Summary

| Article | Topic                                | MeshAI Feature                                         | Coverage |
| ------- | ------------------------------------ | ------------------------------------------------------ | -------- |
| Art. 6  | Classification of high-risk AI       | [Risk Classification](/governance/risk-classification) | Full     |
| Art. 12 | Record-keeping                       | [Audit Trail](/governance/audit-trail)                 | Full     |
| Art. 13 | Transparency                         | Transparency Cards                                     | Full     |
| Art. 14 | Human oversight                      | [HITL Approvals](/governance/approvals)                | Full     |
| Art. 26 | Deployer obligations                 | Agent Registry + Governance                            | Full     |
| Art. 27 | Fundamental rights impact assessment | [FRIA](/governance/fria)                               | Full     |
| Art. 50 | Transparency for certain AI systems  | Transparency Cards                                     | Full     |
| Art. 73 | Serious incident reporting           | [Incident Reporting](/governance/incidents)            | Full     |

## Article 6 - Classification of High-Risk AI Systems

**Requirement**: Organizations must classify AI systems based on their risk level per Annex III categories.

**MeshAI coverage**: The [Risk Classification](/governance/risk-classification) feature lets you assign risk levels (minimal, limited, high, unacceptable) to each agent. AI-assisted suggestions analyze agent metadata against Annex III categories.

## Article 12 - Record-Keeping

**Requirement**: High-risk AI systems must have automatic logging capabilities that record events throughout the system's lifecycle.

**MeshAI coverage**: The [Audit Trail](/governance/audit-trail) captures core governance actions as immutable events - agent registration/updates/deletion, kill-switch blocks, risk classification, policy changes, and anomaly acknowledgement/resolution. Events are timestamped, attributed to an actor, and exportable in CSV/JSON. (Some actions - approvals, incidents, FRIA, quarantine, ABAC, lifecycle - do not yet emit their own audit event; see each feature's page.)

## Article 13 - Transparency and Information to Deployers

**Requirement**: High-risk AI systems must be designed to be sufficiently transparent to enable deployers to interpret outputs and use them appropriately.

**MeshAI coverage**: Transparency Cards are auto-generated for each agent, documenting purpose, capabilities, limitations, model provider, and risk classification. Access via `GET /agents/{id}/transparency-card`.

## Article 14 - Human Oversight

**Requirement**: High-risk AI systems must be designed to allow effective human oversight during use.

**MeshAI coverage**: [HITL Approvals](/governance/approvals) - via the `require_approval` policy type - inline-enforce human review before agents can execute certain actions. `require_human_review` is a periodic-review flag: it is recorded as evidence, not enforced inline, so it documents a review cadence rather than blocking an action in real time.

## Article 26 - Obligations of Deployers

**Requirement**: Deployers must implement appropriate technical and organizational measures, monitor AI system operation, and keep logs.

**MeshAI coverage**: The Agent Registry provides a complete inventory of all deployed AI agents. Governance policies enforce organizational rules. Real-time monitoring detects anomalies. All actions are logged in the audit trail.

## Article 27 - Fundamental Rights Impact Assessment

**Requirement**: Deployers of high-risk AI must conduct a fundamental rights impact assessment (FRIA) before deployment.

**MeshAI coverage**: The [FRIA feature](/governance/fria) provides structured templates covering all six required assessment areas (a–f). Each FRIA is immutable once created - reassessing an agent creates a new, incrementing version rather than editing the prior one.

## Article 50 - Transparency Obligations for Certain AI Systems

**Requirement**: Providers of AI systems that interact with natural persons must ensure users are informed they are interacting with AI.

**MeshAI coverage**: Transparency Cards document agent purpose and interaction patterns. The `limited` risk classification automatically flags agents that require transparency disclosures.

## Article 73 - Reporting of Serious Incidents

**Requirement**: Providers and deployers must report serious incidents to the relevant market surveillance authority within strict deadlines.

**MeshAI coverage**: The [Incident Reporting](/governance/incidents) feature provides structured incident creation, automatic deadline tracking (15-day and 2-day timelines), and authority notification support.

## Readiness Score

MeshAI calculates a [readiness score](/governance/readiness-score) (0–120) across 7 components that map to these articles. Use it to track your compliance progress and identify gaps ahead of the Annex III (2 December 2027) and Annex I (August 2028) deadlines.