RAG with citations · Verified internal retrieval
RAG with citations, traced to the official source.
RAG with citations returns each ranked passage with an official source anchor and traceable source identity. Follow one measured EU AI Act query through enforced filters, corpus version, passages, EUR-Lex citations, and lineage checksums.
- Status
- Internal reference workload
- Customer access
- Unavailable
- Corpus
- Frozen and source-verified
- Retrieval gate
- Passed · 2026-08-21
Reading the evidence
What to check in a cited passage.
Start with the authority and canonical source URL, then follow the article or annex anchor to the cited text. Check that the language, jurisdiction, and corpus version match the question you are trying to answer.
This trace is frozen at version 2026-08-21.1 and was reviewed on 2026-08-21. Its checksums link the passage, normalized document, and source snapshot to specific artifacts. They do not prove that a source is current or that an application’s final answer is correct.
The historical request retrieved ten results; this page displays five reviewed excerpts. The prepared private-beta plan allows up to five results per request. Use the current request example for integration and the evaluation worksheet to define your own acceptance gate.
What the citation adds
A passage, an official anchor, and the identity of the source that served it.
CorpusMesh returns retrieval context for a customer-controlled model. Each result keeps its citation label, EUR-Lex anchor, corpus version, verification scope, and lineage evidence. The service does not generate the final answer.
The question enters as text, not as a client-made vector.
How do Article 6 and Annex III work together to identify high-risk AI systems?
- Knowledge base
- eu-ai-act-reference
- Requested filters
- language: en · jurisdiction: eu
- Result limit
- top_k: 10
- Profile
- gemini-embedding-2-768-search-v2
Mandatory scope
The request cannot widen its own permissions.
CorpusMesh pins the corpus version, verified state, jurisdiction, language, and authorization class before retrieval.
- Corpus version
- 2026-08-21.1
- Verification
- verified
- Jurisdiction
- EU
- Language
- EN
- Authorization
- internal-preview
The evidence arrives ranked and source-linked.
Scores are observed retrieval scores, not confidence percentages. Open a result to inspect its bounded passage and citation.
2. In addition to the high-risk AI systems referred to in paragraph 1, AI systems referred to in Annex III shall be considered to be high-risk.
- Official document
- Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence
- Citation
- Article 6 Classification rules for high-risk AI systems
- Published / effective
- 2024-07-12 / 2024-08-01
- Chunk identity
eeed91ee4bc5…b4e4b3d7
1. Irrespective of whether an AI system is placed on the market or put into service independently of the products referred to in points (a) and (b), that AI system shall be considered to be high-risk where both of the following conditions are fulfilled: (a) the AI system is intended to be used as a safety component of a product, or the AI system is itself a product, covered by the Union harmonisation legislation listed in Annex I; (b) the product whose safety component pursuant to point (a) is the AI system, or the AI system itself as a product, is required to undergo a third-party conformity assessment, with a view to the placing on the market or the putting into service of that product pursuant to the Union harmonisation legislation listed in Annex I.
- Official document
- Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence
- Citation
- Article 6 Classification rules for high-risk AI systems
- Published / effective
- 2024-07-12 / 2024-08-01
- Chunk identity
00b56a3a4b45…a74b180d
High-risk AI systems pursuant to Article 6(2) are the AI systems listed in any of the following areas:
- Official document
- Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence
- Citation
- Annex III
- Published / effective
- 2024-07-12 / 2024-08-01
- Chunk identity
5240ae49a099…4ef90c1a
High-risk AI systems referred to in Article 6(2)
- Official document
- Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence
- Citation
- Annex III
- Published / effective
- 2024-07-12 / 2024-08-01
- Chunk identity
dd83054cc9f5…6612f437
3. By derogation from paragraph 2, an AI system referred to in Annex III shall not be considered to be high-risk where it does not pose a significant risk of harm to the health, safety or fundamental rights of natural persons, including by not materially influencing the outcome of decision making. The first subparagraph shall apply where any of the following conditions is fulfilled: (a) the AI system is intended to perform a narrow procedural task; (b) the AI system is intended to improve the result of a previously completed human activity; (c) the AI system is intended to detect decision-making patterns or deviations from prior decision-making patterns and is not meant to replace or influence the previously completed human assessment, without proper human review; or (d) the AI system is i [excerpt truncated]
- Official document
- Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence
- Citation
- Article 6 Classification rules for high-risk AI systems
- Published / effective
- 2024-07-12 / 2024-08-01
- Chunk identity
f859f5a2e435…c94fb3e0
One identity chain, from official source to response.
- Official sourceEUR-Lex
- Source set
5ac47834fdf9…d48dfffa - Chunks
2d73f00da7bd…81e74924 - Projectionpgvector
- Responseproduction runtime endpoint response
- Sealed fixtures
- 80
- Recall@10
- 96.2%
- nDCG@10
- 82.0%
- Boundary violations
- 0
- Dataset
- EU AI Act v1, 80 fixtures
- Scope
- Recall and nDCG: 70 answerable queries
- Version
- 2026-08-21.1
- Reviewed
- 2026-08-21
See how citation precision and retrieval quality were measured in this published 80-fixture v1 artifactReview the RAG API response fields
88821ad972ee9c05ed2d56e6f125baa0000c9f9abc4515f34dd82db3ef61fc17What this does not prove
Retrieval evidence is not a legal answer.
- The retrieval endpoint remains restricted to an internal bearer token.
- The trace returns source context and does not generate or validate a legal answer.
- The internal reference corpus is not an available commercial knowledge base.
Build with bounded knowledge
Need maintained knowledge for your own application?
Tell us which sources, jurisdictions, and update obligations your product depends on. We will use the request to prioritize coverage, not pretend that a planned corpus already exists.