Research EU legislation, Court of Justice case law, and Official Journal publications through AI assistants. Search across 24 official languages, retrieve structured legal documents and historical versions, run semantic searches, and trace citations across a graph of 650K+ legal references.
Antigravity is Google's agentic IDE. Remote MCP servers are added through its raw config using the serverUrl field.
This guide connects LexAPI MCP to Antigravity through MewCP, so Antigravity can use LexAPI MCP 's 33 tools, like lex_info, lex_search and lex_recent_documents, without your LexAPI MCP credentials ever reaching the app.
Your LexAPI MCP API key is stored encrypted and injected only at call time. Your agent never holds it, so it can never leak it, log it or paste it into a prompt.
More apps
Manage courses, assignments, rosters and student activity within Google Classroom. Create learning workflows, track submissions and organize educational content across classrooms.
PubMed enables AI agents to search biomedical and life sciences literature, retrieve research abstracts and publication details, filter studies by author, date, and topic, and explore scientific evidence to support medical research and literature reviews.
Perform advanced mathematical, statistical and symbolic computations. Solve equations, evaluate expressions and support scientific and engineering calculations.
Consensus MCP lets AI assistants search 400+ million peer-reviewed research papers. Connect your preferred client below. You’ll be searching in under 2 minutes
Other clients
LexAPI MCP works the same way in every client below. Pick yours for its own guide.
Free to start.
Scope access to the exact LexAPI MCP tools and accounts you allow, work and personal side by side. Anything outside that scope is refused at call time, even when the model asks for it.
Each call is logged: what ran, on which account, and whether it worked. Failures come back structured, so the agent knows whether to retry instead of guessing.
Every app sits behind one secure MCP URL, and there is no server for you to deploy or keep running. Add an integration in the dashboard and it is live on the next call, with no config to change and no new credentials to wire up. Tools load only when a task needs them, so your agent stays fast and focused however many apps you connect.
In the agent panel click "…" → MCP Servers → Manage MCP Servers → View raw config.
Remote servers use the "serverUrl" field.
Save the file. Sign in when the browser window opens.
Start a new chat and ask: "List the tools you can use from MewCP." You should see your connected apps.
The caller's LexAPI plan, credit balance (remaining, reset date) and limits, plus the accepted values for search filters. Free (0 credits). Use it when the user asks about their plan or credits, or to check the connection.
Structured EUR-Lex search: keywords plus filters (dates, type, author, procedure, collection) → CELEX numbers, titles, dates; read a hit with lex_get_document. Matching is literal: use 1–3 distinctive terms; for concept questions use lex_semantic_legislation / lex_semantic_case_law. Cost: 3 credits per result page (~10 results; unfetched pages refunded).
Documents published on EUR-Lex in the last N days (default 7), newest first, optionally filtered. Use for "what was published this week". Cost: 3 credits per 10 results of limit (clamped to the plan); an unfiltered request is answered from the Publications Office in one query for 3 credits.
Turn a pasted EUR-Lex URL, an ELI URI, an ECLI (e.g. ECLI:EU:C:2020:559) or a bare CELEX into the canonical CELEX number plus EUR-Lex links. Cost: 1 credit. ECLIs resolve only for case law already in the LexAPI corpus (404 otherwise). resolvedVia says how: celex / url (read off the identifier), ecli-corpus / eli-corpus (corpus lookup), eli-cellar (Publications Office register), eli (derived from the ELI pattern only). verified: false is normal for CELEX and URL inputs (they are not looked up, so a mistyped CELEX also comes back); with resolvedVia "eli" the mapping itself is a guess — check it with lex_get_document before relying on it.
A parsed EU document by CELEX: metadata plus body (articles, recitals, judgment paragraphs, annexes). Large documents return an outline first; then read an article with articleId or parts with include. content: null means no published text — textAvailability says why and may point to the latest consolidated version. Consolidated CELEXes (0…-YYYYMMDD) are labelled: not the authentic text, cite the base act. Cost: 1 credit from the corpus, 4 via the Publications Office, 5 live (every call is charged).
Fetch up to 50 documents by CELEX in one call (batch size is clamped to the plan: FREE 1, STARTER 5, PROFESSIONAL 20, higher 50). Per-document failures are listed in errors; documents without published text stay in documents with textAvailability. Same per-document cost as lex_get_document (1 / 4 / 5 credits); trimmed items are not charged.
Metadata only for a CELEX — title, type, author, dates, ECLI/ELI, EuroVoc keywords, EUR-Lex link — without the body. Use to identify a document cheaply before deciding to read it. Cost: 1 credit from the corpus, 4 via the Publications Office, 5 live.
Same as lex_get_document, for a eur-lex.europa.eu URL the user pasted (the CELEX is extracted from CELEX:… or uri=…). Always English. For ELI/ECLI links use lex_resolve first. Cost: 1 / 4 / 5 credits by source.
Every text change LexAPI observed on a document, newest first. These are snapshots of the document as LexAPI fetched it — NOT the law in force on a date (for that use lex_semantic_legislation with asOf, or a consolidated CELEX). History starts when the document entered the corpus. Cost: 1 credit.
The text of one observed snapshot — by version number, or the snapshot LexAPI held on a date (YYYY-MM-DD). This is what LexAPI had fetched then, NOT the law in force on that date; dates before the document entered the corpus return 404. Cost: 1 credit.
Find EU legislation articles by meaning. Hits: law_id (CELEX), article_ref (use as articleId in lex_get_document), law_title, score, snippet. Consolidated hits (consolidated: true) are not authentic: quote them, cite baseCelex. asOf gives the text in force on a past date. Cost: 5 credits; with asOf, plus up to 5 per in-force version fetched live (default max 3 per call; failed fetches free).
Find CJEU / General Court judgments, orders and AG opinions by meaning. Returns cases with celex (feed to lex_get_document), case number, ECLI, court, date, score and a summary. For short keyword-style queries set hyde: true (better recall, 15 credits instead of 5, +1–3 s; refunded to 5 if it falls back). Cost: 5 credits.
Countries with national-law coverage: live status, legislation/case-law coverage (counts, dates, courts), semantic-search availability, citation-graph measurements (citationGraph: methods, rates, floors; null = not measured), and sources with licence/attribution. Call this before any other national tool. Free.
Browse a country's national documents by title/type, newest-first by document date. documentKind ("act" | "decision") tells legislation from case law. Not yet live 404s — check lex_national_coverage first. Cost: 3 credits flat, not per page; counts against the national-law allowance on lower plans (see allowance). Paging: page reaches only the first 1,000 rows; beyond, pass the previous nextCursor as cursor (any depth), without page.
A national document by id, e.g. "AT-NOR-12345". Same fields in every country: documentKind ("act" | "decision"); decisions add court, chamber, caseNumbers, citedNorms; articles {number, title, section, text, inForceFrom, inForceTo} (text: null = no text for that unit). Large documents return an outline first; read on with view "full" and offset. content: null = the source publishes no machine-readable text (textAvailability says why). A document its source withdrew is a distinct error: nothing of it is served; do not retry, drop any copy held. Not yet live 404s — check lex_national_coverage first. Cost: 5 reserved, settles at 1 (an unknown id too), so a pool under 5 402s. Counts against the national-law allowance on lower plans (see allowance). Show attribution (source/licence) to the user.
Find one country's case law by meaning. Hits carry documentKind, ecli, dateOfDocument, and court/caseNumbers where indexed. hyde is a no-op here (no rewrite, no extra cost). Not yet live errors — check lex_national_coverage first. partial: true (partialReason SEARCH_INCOMPLETE) means the index could not complete the search: results may be incomplete — say so, and follow the hint. Cost: flat 5 credits, never 15; counts against the national-law allowance on lower plans (see allowance). Each hit's attribution names the source/licence — show it to the user.
Find one country's legislation articles by meaning. No asOf — current text only. Not yet live errors — check lex_national_coverage first. partial: true (partialReason SEARCH_INCOMPLETE) means the index could not complete the search: results may be incomplete — say so, and follow the hint. Cost: flat 5 credits; counts against the national-law allowance on lower plans (see allowance). Each hit's attribution names the source/licence — show it to the user.
Inbound citations: documents that cite, amend, correct, repeal, implement, transpose or are based on this act, with the edge type and its provenance. For legislation, amendment / corrigendum / repeal edges are read from the Publications Office (Cellar) and refreshed on read when older than 7 days; if that refresh does not finish in ~2.5 s, relationSync.status is "pending" (or "skipped") and the stored edges are served — check relationSync before calling the list complete. Other edge types cover only documents LexAPI has extracted. Page until offset + limit >= totalEdges. includeNational adds national neighbours of live countries (BUSINESS+ only; also accepts a national id, e.g. "AT-NOR-12345"). Cost: 1 credit.
Outbound citations: what this document references, amends, repeals, implements, transposes or is based on, with the edge type and its provenance. Only populated for documents whose citations were extracted (lex_extract_citations). includeNational adds national neighbours of live countries (BUSINESS+ only; also accepts a national id, e.g. "AT-NOR-12345"). Cost: 1 credit.
Both directions at once — documents this one cites and documents citing it — deduplicated, with edge types and provenance per neighbour; paging.citesTotal / citedByTotal give the full counts. includeNational adds national neighbours of live countries (BUSINESS+ only; also accepts a national id). Cost: 20 credits, refunded on a partial (timeout) answer (prefer lex_cited_by / lex_cites at 1 credit when one direction is enough).
Shortest chain of outbound citations from one document to another (e.g. how a judgment reaches a directive), optionally over filtered edges only. found: false is a normal answer, not an error. includeNational follows national documents of live countries too (BUSINESS+ only; from/to also accept a national id). Cost: 20 credits.
Documents that cite the most of the same acts as this one (bibliographic coupling) — "what else deals with this"; the edge filters narrow which shared citations count. Needs the seed's outbound citations (lex_extract_citations). includeNational considers national documents of live countries too (BUSINESS+ only; also accepts a national id). Cost: 20 credits.
Totals for the whole LexAPI citation graph plus the 10 most-cited and 10 most-citing documents. includeNational counts national documents of live countries too and adds byJurisdiction, byProvenance and transposition (notified measures, and how many LexAPI holds) — BUSINESS+ only. Cost: 20 credits.
The measures member states (and the UK) notified to the Commission for an EU act (CELEX sector 3, e.g. a directive), by country, from the Publications Office's Cellar records. A notification is not a finding that the measure transposes the act, or transposes it completely — say so when presenting it (the result's note). Every plan. On BUSINESS+ it also says, for live countries only, which measures LexAPI holds (id, resolved) and adds documents whose national source records the relation (provenance "source"); below that those fields are absent. termsRequired: true = the account must accept the national-source terms in the dashboard first. Keep source.cellar.attribution with anything republished. Output is bounded (limit, ~40,000 characters): page with the offset given in more. Cost: 5 credits per call (each offset page too); not an EU act = 404, settles at 1.
A national document's counterparts in the other live countries, through the EU acts it transposes or implements (via), ranked: 1 = the national source records the relation; 2 = a notified measure LexAPI holds; 3 = a notified Cellar record LexAPI does not hold (no id). An EU act (CELEX sector 3) as id groups its measures the same way. Only live countries are returned — check lex_national_coverage; a national id of a country not live 404s (nothing charged); any other 404 settles at 1 credit. A notification is not a finding that the measure transposes the act. BUSINESS, SCALE and ENTERPRISE only (403 below, refunded); the national-law allowance does not apply. Output is bounded (limit per act, up to 25 acts and ~40,000 characters): page with the offset given in more. Cost: 20 credits (each offset page too); a server timeout is refunded — narrow with jurisdictions.
Read a document's EUR-Lex notice and add its outbound citations to the graph, so lex_cites / lex_related work for it. Idempotent: returns alreadyExtracted when done before. Cost: 5 credits (1 when already extracted).
Rows from the LexAPI corpus matching filters — e.g. regulations dated 2024 or documents whose text changed since a date. BUSINESS plans and above (403 otherwise). Cost: 1 credit per row returned, at least 1 per call. The call is admitted for 1 credit, the matching rows are counted, and the rest of 1 credit per row it can return (up to limit) is reserved; if the pool cannot cover that, LexAPI answers 402 before returning anything (retry with a smaller limit). Reserved credits for rows not returned are refunded. Keep limit small. An export that fails part-way is reported as incomplete, never as a result. Bulk export for ML training needs a separate licence (Terms §6).
The account's webhooks with status, criteria, watch list and delivery counters. Free.
One webhook with its recent delivery attempts (status, HTTP response, attempts). Pass deliveriesOffset/deliveriesLimit to page through older deliveries. Free.
Subscribe a URL (must be a public https URL) to (a) new EUR-Lex documents matching searchCriteria and/or (b) changes to watched acts (watchCelex: amendments, corrigenda, repeals, new consolidated versions). Checked hourly; deliveries are free and HMAC-signed. The signing secret is returned ONLY in this response — tell the user to store it. Plan limits apply (FREE has no webhooks). Confirm the URL and criteria with the user first. Free.
Change a webhook's name, URL, searchCriteria (null removes them), watch list (replaces the whole list) or status. status ACTIVE re-enables a PAUSED or FAILED webhook and re-queues its recent failed deliveries; PAUSED stops deliveries. Free.
Permanently delete a webhook and its delivery history. Cannot be undone — confirm with the user first. Free.
POST a signed webhook.test payload to the webhook URL now and report the endpoint's response. Use to check an endpoint is reachable and verifies signatures. Free.