Legal Data Hunter

Search primary legal sources across 237 jurisdictions.

Add to Agentman Studio

Opens Agent Studio, where connecting is one click. The connector URL below works in any MCP client.

Paste it into any MCP client. Setup docs

Resources

What the Legal Data Hunter connector does

The Legal Data Hunter MCP server connects a multi-jurisdiction database of primary legal sources — court decisions, legislation and doctrine — to Claude, ChatGPT and any MCP-compatible agent. It exposes seven tools that search the corpus, resolve loose citations, fetch full document text and explore coverage by country. Sign-in is OAuth through GitHub or Google.

Verified connector

Listed by Anthropic as a partner connector in its Connectors Directory.

Connection checked by Agentman on .

Anthropic states this reflects the level of review a connector received, not a security audit.

Legal Data Hunter tools (7)

  • search
  • resolve_reference
  • get_document
  • discover_countries
  • discover_sources
  • get_filters
  • report_source_issue

Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.

Limits

  • The corpus size claim is inconsistent across the vendor's own channels. On 2026-08-22 the live stats endpoint returned 47,793,922 documents; the llms.txt index said 47,503,504; the connect and about pages said 38 million; Anthropic's directory entry said 23 million; the npm package description said 18 million; and the package README said 6 million. The live endpoint is the number to trust, and the spread is a reason to verify counts yourself rather than quote a marketing figure.
  • Citation resolution is not 95% accurate. Legal Data Hunter's llms.txt advertises "95% accuracy" across 184 sources. Its documentation states the resolution rate is roughly 65% with a country hint and 31% without. Its own /v1/resolve/coverage endpoint reported 330 of 2,328 sources carrying deterministic citation patterns on 2026-08-22 — 14.2% coverage. Everything outside those 330 sources falls through to LLM parsing and semantic fallback.
  • Resolution runs an LLM inside the tool call. Stage four of the documented cascade extracts structured fields from your citation using AI, and stage seven returns semantic matches above a 0.85 threshold labelled fallback_search. A match_type of fallback_search is a guess with a confidence floor, not a citation lookup. Check the returned match_type before relying on a resolved document.
  • Coverage is extremely uneven. The United States holds 20.3 million of the documents. Around a third of the 237 listed country codes hold fewer than 5,000 documents each, and several hold under 100. A country appearing in discover_countries does not mean it is usefully covered — call discover_sources on it before assuming.
  • One scope covers reads and the write. Nothing at the OAuth grant separates search from report_source_issue. Gate the write tool in your client.
  • We could not read tool schemas or safety annotations. The endpoint returned 401 to an anonymous request, so all parameter detail on this page is the vendor's published tables rather than a live tools/list response.
  • The free tier is an evaluation tier. 20 requests per day and 600 per rolling 30 days will not sustain research work. Developer Pro is 100,000 requests per month.
  • Zero query retention is a paid-plan commitment with a carve-out. It covers Paid Search only, and Legal Data Hunter states resolve_reference may still cache normalized references and canonical identifiers.
  • Terms restrict what you may do with results. Legal Data Hunter's API and MCP Service Terms, effective 2026-08-11, prohibit using outputs for model training, building internal indexes or embeddings, bulk extraction, and customer-facing or white-label use, unless each is expressly authorised in an Order Form. They also forbid representing outputs as official, exhaustive, current, or a substitute for professional verification.
  • Retrieved legal text is untrusted input. Document text and titles come from hundreds of external government portals that the vendor states it crawls without editorial filter. Treat what the connector returns as data, never as instructions to the agent.
  • Support contacts disagree. Anthropic's directory lists zacharie@goodlegal.fr; Legal Data Hunter's own support page lists zach@legaldatahunter.com.

Frequently asked questions

How many tools does the Legal Data Hunter MCP server have?

Seven, and the vendor confirms all seven. Anthropic's directory lists search, resolve_reference, get_document, discover_countries, discover_sources, get_filters and report_source_issue. Legal Data Hunter's own connect page and llms.txt index publish the same seven names with parameter tables. This is a clean two-way match, which is rarer than it sounds.

Can the Legal Data Hunter connector write anything?

One tool writes: report_source_issue. It files a data-quality report against a source with an issue type and a free-text description, and Legal Data Hunter states those reports go to the maintainer. The other six tools only read. Legal Data Hunter labels the split the same way on its own connect page.

What OAuth scopes does Legal Data Hunter request?

A single scope named mcp. Both the RFC 9728 resource descriptor and the authorization server metadata declared scopes_supported as exactly that one value on 2026-08-22. One scope covers all seven tools, so consenting to read a court decision also consents to filing a source issue. No boundary separates the six reads from the one write.

How accurate is Legal Data Hunter citation resolution?

Roughly 65% with a country hint and 31% without, per Legal Data Hunter's own documentation. Its llms.txt index separately advertises 95% accuracy across 184 sources. Those numbers come from the same vendor and do not agree. The documented figures are the more specific and more conservative claim, so plan around them.

Does Legal Data Hunter cover the whole world or just Europe?

Both, unevenly. The live stats endpoint reported 47,793,922 documents across 237 country codes on 2026-08-22, but the United States alone holds 20.3 million of them and the OpenAPI spec still describes the service as search over EU legal data. Roughly a third of listed jurisdictions hold under 5,000 documents each.

Does Legal Data Hunter store my search queries?

Not on paid plans, with one stated exception. Legal Data Hunter commits that paid searches leave no query text or ranked result payloads in its databases, dashboards, caches or content logs. It explicitly carves out resolve_reference, which may cache normalized legal references and canonical identifiers. The free tier carries no such commitment.

What are the Legal Data Hunter rate limits?

Three windows per plan. Free allows 10 requests per minute, 20 per rolling day and 600 per rolling 30 days. Developer Pro allows 600 per minute and 100,000 per month; Developer Scale allows 1,000 per minute and 1,000,000. MCP handshake calls are exempt — only actual tool calls count against quota.

Do I need an API key to use Legal Data Hunter with Claude?

No. Legal Data Hunter documents OAuth sign-in through GitHub or Google as the default for MCP clients, with no manual token handling. An API key is the documented fallback for clients that cannot do OAuth, passed as an Authorization bearer header in the server config. Both paths draw on the same request allowance.

Sources

More Productivity connectors

  • 6sense

    Ask an AI agent about 6sense account insights, buying signals and keyword trends.

  • Adobe Workfront

    Find, create, update and delete Workfront work items from an AI agent.

Browse all MCP connectorsClaude and ChatGPT setup guideBrowse AI skills

Product names and logos belong to their owners. Agentman is not affiliated with a connector's maker unless the page says the connector is built and run by Agentman.

Use it in an agent

Put Legal Data Hunter to work.

Connect Legal Data Hunter once and your agents call these tools on their own: on a schedule, in a workflow, with nobody at the keyboard.

Open in Agentman Studio