Ris Austria MCP Server

@cyanheads/ris-austria-mcp-server Search Austrian consolidated federal & state law, court decisions, and the authentic Bundesgesetzblatt from the official RIS via MCP. Keyless. STDIO or Streamable HTTP. 9 Tools • 1 Resource

Public Hosted Server: https://ris-austria.caseyjhand.com/mcp

Overview

Austrian law from RIS, the government's official legal information system: federal, state, and municipal law, court decisions, official gazettes back to 1848, draft legislation, and ministerial decrees. Every document is labeled with its binding status. No API key required.

Tools

Resources

The resource mirrors ris_get_document, so tool-only clients lose nothing.

Capability reference

ris_search_legislation tool

• Search by query (full text) or title (title, short title, or abbreviation such as "DSG"). scope is federal (default) or a Bundesland, municipality switches a state scope to municipal law, and language: english searches the ~138 English translations
• in_force_as_of defaults to today in Austria and is echoed as appliedInForceAsOf; include_all_versions and the entered_force_* / left_force_* windows are mutually exclusive alternatives
• Records carry law_id (filter on it for every section of a law), section_label, abbreviation, keywords, in-force dates, eli, celex_references, and content_urls

ris_search_case_law tool

• court is required, one code per call. Filter by query, cited norm ("DSG §1"), case_number, decided_from/decided_to, and decision_type (headnote, full_text, or all, the default). Court-specific filters include court_name (justiz), state (lvwg/uvs), and collection_number (vfgh/vwgh/uvs); one sent to the wrong court fails as court_filter_mismatch
• Records carry case_numbers (a headnote lists every decision it covers), decision_date, ecli, norms_cited, and guiding_principle on headnotes
• court: normenliste returns indexed laws under norm_index instead of decisions

ris_search_gazette tool

• scope is federal (default), a Bundesland, district, or municipal. Filter by number, part, type, published_from/published_to, issuer, district_authority, or municipality; state scopes also take series and state_era
• Federal queries are served from one era (BgblAuth 2004+, BgblPdf 1945–2003, BgblAlt 1848–1940), echoed as servedApplication; a date range spanning two eras fails as cross_tier_range with the dates to split at
• Records carry binding (authentic, historical_record, consolidated_informational), authentic_pdf_url where one exists, alex_url (ÖNB scan) for 1848–1940, celex_references where applicable, and promulgation on federal entries from 1848–2003

ris_search_drafts tool

• stage is required: review_drafts (with in_review_on) or government_bills (2004+, with decided_from/decided_to); ministry accepts an abbreviation such as "BMF" or a full designation
• Records carry review_start and review_deadline or decided, abbreviation and keywords where RIS supplies them, plus materials[] (Erläuterungen, Textgegenüberstellung, Vorblatt/WFA, annexes) classified by type. Pass a materials[].url to ris_get_document to read one

ris_search_announcements tool

• collection is required: social_insurance, veterinary, court_rules, trade_exam_rules, health_structure_plans, ministerial_decrees, or council_minutes. A filter its collection doesn't accept fails as collection_filter_mismatch, including changed_since on social_insurance and veterinary. issuer expands ministry ("BMF") and social-insurance carrier ("ÖGK") abbreviations
• Records carry binding (authentic, administrative_directive, or preparatory), authentic_pdf_url, and document_url, the only browsable view of council minutes and decrees
• Each collection adds its own metadata where RIS supplies it, such as case_numbers, in_force_from/in_force_until, keywords, court_name, plan_state, and session_number

ris_lookup_citation tool

• One citation per call, classified as a norm ("§ 6 DSG"), gazette number ("BGBl. I Nr. 165/1999"), case number ("VfGH G 13/2025") or ECLI, or collection number ("VfSlg 19.632/2012"). kind forces a route, court and state narrow it, and in_force_as_of pins a norm's version (default today)
• Returns found, kind, and a record in the matching search tool's shape, with alternatives_count when several documents matched. A miss returns found: false with guidance instead of an error
• A norm resolves to the cited law:
• the law is matched as RIS files it ("§ 2 EStG" → EStG 1988 § 2; "AlVG 1977" → the AlVG, filed without its year; "VwGVG 2014 §28" → VwGVG § 28), and a long title is read by its short form ("StaatsanwaltschaftsG - StAG §35" → StAG § 35)
• sub-provisions and a trailing ";" are set aside ("Art. 6 (1) B-VG" → B-VG Art 6), and an ordinal-led abbreviation reads like any other ("1. COVID-19-JuBG §1")
• a cited version is set aside ("ABGB idF BGBl I 2010/28 §988" → ABGB § 988 as in force on the requested date, never the amending act's gazette entry), with a note saying the cited version was not applied
• an entry citing several provisions ("DSG §1, §24", "ZPO §179. ZPO §275 Abs2") resolves the first and names the rest
• An EU act RIS does not consolidate ("DSGVO Art32", "RL 2006/123/EG", "VO (EG) 178/2002 Art3") returns found: false with its EUR-Lex link, without a RIS request
• A gazette number without its part ("BGBl. Nr. 50/2025") comes back ambiguous, naming parts I, II, and III

ris_get_document tool

• Address by document_number + application, or by a document_url from content_urls or a draft's materials[].url; format is markdown (default), html, xml, or urls_only
• Every result carries binding_status and authentic_pdf_url where one exists. Where a PDF is the only rendition (district and municipal promulgations, court rules, party-transparency decisions, council minutes, and draft companions filed without HTML), markdown returns the text extracted from it: plain text with a --- Page N of M --- line before each page, text_source: pdf, and a notice naming the PDF. A scanned PDF with no text layer, one over 16 MiB, or one that cannot be fetched in time returns a notice with the usable URL instead of failing, as do html and xml there and the 1848–1940 gazettes, which have no PDF to read
• Markdown writes image references as absolute URLs on the content host. A rendition that is nothing but page images reads its PDF's text, and keeps the image URLs with a notice when the PDF has none
• Markdown over 40,000 bytes returns kind: outline; re-call with sections to fetch entries. html and xml over 40,000 bytes return kind: link with byte_size and content_urls, so read with markdown

ris_track_changes tool

• application (any of the 39 codes) plus changed_from/changed_to; include_deleted adds removals, which no other tool reports
• Records carry changed, published, and binding_status; deletions set deleted: true with deleted_at
• Law rows carry section_label, law_id, and in_force_from/in_force_until (no end date is not the same as in force now), and decision rows carry case_numbers and decision_date

ris_list_reference tool

• topic selects one of 17 static tables: applications, courts, states, decision types and kinds, issuing bodies, ministries, collections, stages, changed-since intervals, section types, gazette parts, law types, district authorities, justiz subject areas, search syntax, and citation formats
• Returns summary and entries (value, label, details) with no upstream call

ris://document/{application}/{documentNumber} resource

• text/markdown of one document, addressed by application + documentNumber from a search or lookup result; same content as ris_get_document with format: markdown
• Applications that publish only a PDF return its extracted text in a fenced block with a note naming the PDF; the 1848–1940 gazettes and a PDF with no extractable text return a note pointing at the PDF or scan. Oversized text returns the outline, so use the tool to fetch