Loading ToolBench
MCP server for healthcare AI — medical coding, DMEPOS fee schedules, drug intelligence, reimbursement rates, clinical trials, physician payments, disease surveillance, prior authorization, NER, claims validation, HIPAA compliance. 32 tools — 6 free no-auth + 24 pay-per-call via credits or USDC + 2 account tools — plus DME billing prompts and CMS dataset resources.
MyMedi-AI MCP server demonstrates solid definition quality across 12 healthcare-focused tools. Strengths: all tools have verb-noun naming conventions (pa_required_check, code_lookup, denial_code_info); descriptions are detailed and domain-specific (100-250 chars average), clearly explaining when and why to use each tool; parameters uniformly use Zod schema with descriptions; most tools document their medical/compliance context. Weaknesses: descriptions are verbose (some exceed 300 chars, wasting tokens); a few parameter descriptions lack actionable constraints (e.g., modifier_advisor's 'scenario' parameter accepts free text without guidance on expected patterns); output schemas are not explicitly documented in the tool definitions (medical coding tools like code_lookup return complex fee-schedule objects but the response structure is inferred, not declared); limited error recovery guidance (no explicit documentation of how to handle API failures, rate limits, or invalid code formats); no batch tools include per-item success/failure reporting. Security and composition are sound, all tools are read-only with proper annotations, and tools chain naturally (code_suggest → code_lookup → reimbursement_basic flow). Tool density is appropriate (12 tools, none combining multiple concerns).
Look up a medical code (ICD-10, CPT, HCPCS). Returns description, category, active status, and related codes. For DMEPOS (HCPCS) codes it also returns a labeled fee schedule: per-modifier (RR/NU/UE) national min–max ranges, or — when a state is given — that state's exact non-rural and rural rates. Source: CMS DMEPOS Fee Schedule (DME26-B).
Look up basic metadata for a medical code: code, codeType, description, category, isActive. Basic metadata only — the paid code_lookup adds full metadata. Free, no API key required.
Look up a list of medical codes (ICD-10, CPT, HCPCS) in one call. Per-item results mirror code_lookup (description, category, active status, related codes, DMEPOS fee schedule with optional state filter). Priced per code — $0.001 × number of codes, max 25 per call; the full charge is refunded automatically when every code misses.
Suggest ICD-10/CPT/HCPCS codes from a clinical description. Term-based search over the 81K-code CMS database, ranked by matched-term coverage and relevance. Works with natural sentences ("patient with obstructive sleep apnea prescribed CPAP"). Automatically refunds the call when nothing matches.
Output schemas not explicitly documented. Complex tools like code_lookup return nested DMEPOS fee-schedule objects, but the response structure is inferred from the description rather than formally specified. LLMs cannot plan downstream tool calls or extract chaining IDs without declared output types.
Parameter descriptions exceed 250 characters in several tools (modifier_advisor, pa_predict, order_readiness_checklist), diluting signal and wasting tokens. Descriptions should be 50-150 chars, state WHAT the param does, not entire regulatory citations.
modifier_advisor parameter 'scenario' accepts free-form text with no validation guidance. Examples in description ('ABN on file', 'bilateral') may bias LLM to reuse exact phrases instead of adapting to context. Replace with enum or provide explicit format constraints.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | C | 67 | 2026-07-28+ | v2 |
Validate a medical code for correctness, active status, and context. Returns warnings and errors.
Validate a list of medical codes for correctness, active status, and optional date-of-service context in one call. Per-item results mirror code_validate (valid, active, warnings, errors, codeDetails). Priced per code — $0.005 × number of codes, max 25 per call. An invalid code is a billable answer (valid:false), same as the single validate.
Explain a DME claim denial code (CARC). Returns title, meaning, common DME causes, fixes, appealability, and related codes. Free, no API key required.
Editorial guidance on DMEPOS billing modifiers: the KX/GA/GY/GZ medical-necessity and liability family, RR/NU/UE rental-vs-purchase, capped-rental month markers (KH/KI/KJ), and RT/LT laterality. Pass a HCPCS code to scope guidance to that item's DMEPOS category, or a category directly; add a scenario phrase (e.g., "ABN on file", "bilateral") to surface the relevant modifiers. Original editorial content, not payer policy. PHI-free. Free, no API key required.
Blank pre-delivery checklist for a HCPCS DMEPOS code: the universal standard written order (SWO) elements (42 CFR 410.38(d)), whether the code requires a face-to-face encounter and written order prior to delivery (F2F/WOPD), and whether it is on the Medicare Required Prior Authorization List. Requirement definitions only — PHI-free, never send patient data. Free, no API key required.
Prior-authorization outlook for a procedure code. When a historical cohort of decided PAs exists (≥10), returns a data-driven approval rate with cohort size and confidence. When no cohort exists, returns the verifiable facts instead — CMS Required Prior Authorization List status, category, and published review timeframes — and explicitly reports that no probability was computed (the call is refunded on this path). Never fabricates a probability. Original Medicare FFS scope.
Check whether a HCPCS code is on the CMS Required Prior Authorization List (42 CFR 414.234). Returns paRequired flag, category, nationwide-since date, and list version. Original Medicare FFS scope. Free, no API key required.
Look up Medicare payment for a code. Returns the national PFS facility and non-facility payment (CMS RVU × conversion factor) for professional services, plus DMEPOS fee-schedule ranges (rental/purchase, min–max across state fees) for DME items like E/K/L codes. Free, no API key required.
No error recovery guidance. Tools lack descriptions of what to do when invalid HCPCS codes are passed, rate limits are hit, or codes are not found. Error responses should guide LLM toward retry, lookup, or fallback strategies.
Batch tools (code_lookup_batch, code_validate_batch) do not document per-item success/failure. If 1 of 25 codes fails validation, does the entire call fail or return partial results? Unclear handling forces agents to retry all rather than work around individual failures.