Healthcare MCP server
What it does
Three MCP servers for dispatching vendor requests to credentialed service providers across legal, insurance claims, and architecture/engineering/construction verticals. AI assistants can send work requests and receive bids back within ten minutes.
- Dispatch legal work (court reporters, medical exams, e-discovery, translation, expert witnesses) and receive vendor quotes in real time.
- Route insurance claims tasks (medical exams, surveillance, peer review, life-care planning) to qualified vendors with credential verification.
- Coordinate AEC subcontractor dispatch, prequalification checks, bonding capacity verification, and safety records pulls.
Tools this server exposes
| Tool | Does |
|---|---|
| scope_dispatch_matter | Send a matter to vendors and receive live quotes. |
| scope_briefing | Get a daily briefing across all open matters. |
| scope_get_matter | Retrieve details for a specific matter. |
| scope_list_matters | List all matters in the system. |
| scope_list_categories | List available service categories. |
| scope_list_vendors | List available vendors. |
| scope_list_roster | List the current vendor roster. |
| scope_set_vendor_tier | Set the tier level for a vendor. |
| scope_remove_from_roster | Remove a vendor from the roster. |
| scope_award_matter | Award a matter to a selected vendor. |
| scope_roster_audit | Audit the current vendor roster. |
| scope_vendor_health | Check the health and performance status of a vendor. |
| scope_spend_rollup | Get a rollup of spending across vendors. |
| scope_credential_alerts | Receive alerts about vendor credential issues or changes. |
| swp_clarify | Clarify details in an existing negotiation session using Scope Work Protocol. |
| swp_accept | Accept terms in an existing negotiation session using Scope Work Protocol. |
| swp_reject | Reject terms in an existing negotiation session using Scope Work Protocol. |
| swp_session_status | Check the status of a Scope Work Protocol negotiation session. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.