legalos

Endpoint

Hosted endpoint · transport http

Connect

This is a remote MCP server — point an MCP client at the endpoint URL. Clients with native remote support (Claude, Cursor, VS Code, …):

For stdio-only clients (older Claude Desktop, etc.), bridge with mcp-remote:

If the server requires authentication, add the provider's token/header (e.g. an X-API-Key) per its docs. A quick reachability check:

Tool quality observed · not scored

Heaviest: add_beneficiary_intake_entry 2111, set_beneficiary_intake_answers 1715

From an anonymous initialize + tools/list capture measured 2026-10-20; tokens are chars/3.6 of each tool definition. Recorded for the Kin Score's agent readiness as an observed signal — it contributes no points yet.

Tools

• list_cases — List the caller's accessible LegalOS cases with beneficiaryName, canonical status, and a first-party case link. Cursor-paginated. When the user names a case, person, or visa type without an ID, call this proactively and match on beneficiaryName — never ask the user for case IDs or UUIDs.
• get_case_status — Get one case's canonical LegalOS status: waiting party, progress, steps, and server-generated next actions. Never infer beyond it. When listing steps or documents, mark what is done with ✅ and what is still needed with ☑️.
• list_missing_requirements — List a case's outstanding document requirements with status and a first-party upload link. Cursor-paginated. Show items as a ☑️ checklist (✅ when collected). Match the user's document to one, then send them to its requirementUrl — it opens the exact upload row. Never ask for the file in chat.
• get_faq — Vetted LegalOS help content for common client questions. Ground answers ONLY on this content and the user's case data — never give legal advice beyond them. If they do not answer the question, direct the user to the LegalOS support channel (supportChannelUrl).
• list_document_suggestions — Optional, case-strengthening document suggestions for one case: attorney-recommended evidence plus optional tracker items. Never blockers — mention positively after the required upload path is settled, e.g. 'this would also strengthen your case'. May be empty; never invent suggestions beyond it.
• get_beneficiary_intake — Overview of one case's beneficiary intake: per-step completion counts, list totals, the current intakeRevision, and first-party form links. Show the seven steps as a ☑️ checklist (✅ when complete). Read this before any step; submitting stays on the first-party form page.
• get_beneficiary_intake_step — Read one beneficiary intake step: scalar steps return each field's current value and policy; list steps a bounded entry window with the total. All fields, including identifiers (SSN, passport, I-94), return their value and are editable. A truncated list's remainder is on the form page.
• set_beneficiary_intake_answers — Set beneficiary intake answers the user stated — never inferred. Confirm the target case first: verify the beneficiary name; if several cases match, ask — never guess. Echo previousValue + expectedIntakeRevision from a fresh read; re-read on revision_mismatch. Sensitive IDs stay on the form page.
• add_beneficiary_intake_entry — Append ONE user-stated entry to a beneficiary intake list (employment, education, travel/visa, dependents). Confirm the target case first: verify the beneficiary name — never guess. Never edits existing entries (form page only). Needs a fresh expectedIntakeRevision; travel/visa returns entryId null.
• update_beneficiary_intake_entry — Update fields of ONE existing beneficiary intake entry the user identified (employment, education, dependents). Confirm the target case and restate the entry first — never guess. Echo previousFields + a fresh expectedIntakeRevision; re-read on revision_mismatch. Sensitive IDs stay on the form page.
• remove_beneficiary_intake_entry — Remove ONE existing beneficiary intake entry (employment, education, dependents) the user explicitly asked to delete — restate it and confirm the target case first. Soft-deletes; recoverable by staff. Needs a fresh expectedIntakeRevision. Travel/visa entries stay on the form page.
• list_recommenders — List a case's recommendation-letter recommenders — each one's name, connection type, lifecycle status, and draft-reviewed/signed flags. Call get_recommender for a card's full fields.
• get_recommender — Get one recommender's full card — name, LinkedIn URL and background, connection, relationship and project details, note, the draft editor link, status, and a signed-letter upload link. Null fields show what's missing.
• update_recommender — Edit a recommender's card fields or mark its draft reviewed (draftReviewed). Requires expectedRecommenderVersion from get_recommender; a concurrent change is refused. Signed-letter upload stays first-party.
• list_evidence_criteria — List a case's extraordinary-ability evidence criteria: each one's type, status, item counts, and the cached combined-strength verdict. A lean summary — call get_evidence_criterion for one section's items and full gap analysis. May be empty before the section is set up.
• get_evidence_criterion — Get one evidence criterion in full: its items (title, link, processing status, per-item AI evaluation), the combined strength analysis with coverage highlights and remaining gaps, and fulfilled documentation. Uploads happen on the linked first-party page.
• add_evidence_contribution — Add one URL as evidence for a criterion (articleUrl + optional whyItMatters). Additive and async: it returns 'processing' — re-read get_evidence_criterion for the evaluated strength. Document uploads stay on the first-party page.
• send_feedback — Records feedback about this MCP server, including unexpected tool failures, unhelpful responses, or missing capabilities observed by the agent or reported by the user.

About MCP

The Model Context Protocol (MCP) is an open protocol Anthropic introduced for connecting LLM-based agents to external tools and data sources. Providers publish MCP servers that expose their API surface as structured, discoverable tools — an MCP-compatible client (Claude Desktop, Cursor, Cline, Continue, etc.) can connect to the server and call its tools without any per-provider integration code.

Browse every MCP server on the APIs.io network or compare with the broader Agent Skill surfaces of the same providers.

Work with this as data

Every MCP server here is available over the APIs.io API and to AI agents over MCP.

curl "https://apis.io/api/v1/mcp/legalos"

curl "https://apis.io/api/v1/mcp?limit=25"