DenialPath

Deterministic CMS NCCI/MUE claim scrubbing and CARC/RARC denial explanations, cited to source.

CommunityOnlineNo sign-inGlobalFreeRead-only

What it can do

    What data it sees

    Do you need an account

    No: the server works without sign-in

    Deterministic CMS NCCI/MUE claim scrubbing and CARC/RARC denial explanations, cited to source.

    How to connect

    How to connect DenialPath to Claude

    The server address to paste into Claude settings:

    https://denialpath.vercel.app/api/mcp

    1. Open Claude (claude.ai or the desktop app).
    2. Go to Settings → Connectors.
    3. Click Add custom connector.
    4. Paste the server address copied below into Remote MCP server URL and click Add.
    5. In a chat, click + → Connectors and switch the new connector on.

    Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.

    Authorization

    After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.

    How to connect DenialPath to ChatGPT

    The server address to paste into ChatGPT settings:

    https://denialpath.vercel.app/api/mcp

    1. Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
    2. Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
    3. Open Settings → Connectors and click Create.
    4. Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
    5. Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
    6. In a new chat open + → Apps/Connectors and enable the connector.

    OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.

    Authorization

    On first use ChatGPT opens the service's sign-in window. Sign in and approve access.

    How to connect DenialPath to Cursor

    The server address to paste into Cursor settings:

    https://denialpath.vercel.app/api/mcp

    Fastest: click Open in Cursor below and confirm the prompt.

    Manually:

    1. In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
    2. Paste the JSON copied below into ~/.cursor/mcp.json (per project: .cursor/mcp.json in the repo root).
    3. Save the file. The server appears in the MCP list; authorize it there if asked.

    Authorization

    If the service needs sign-in, an authorize button appears next to the server in the MCP list.

    Server tool list (6)

    Raw names from tools/list. Only developers need these.

    check_claimScrub a claim against the NCCI procedure-to-procedure (PTP) edits and medically unlikely edits (MUE) in force on its date of service. Deterministic: no model, so identical input always returns an identical verdict. Every finding.status is one of pass (checked against a real edit and clean), fail (an edit fires), bypassed (an edit fires but a modifier already on the line legitimately resolves it), or no_data (we hold no edit for this code or pair in our dataset). no_data is NOT a pass: treat it as unchecked, since CMS may still deny the line under an edit outside our loaded subset. summary.clean is true only when every finding is pass and none are no_data - never report a claim clean because nothing failed if some lines were never checked. The response echoes editQuarter (the CMS quarter version the verdict was computed against) and coverage (how many PTP pairs and MUE codes are loaded), so you can tell a caller exactly what was checked. Free, no API key needed; a key only raises the rate limit.
    explain_denialExplain a CARC (Claim Adjustment Reason Code) and any RARCs (Remittance Advice Remark Codes) from a remittance: plain meaning, ranked corrective actions, and whether an appeal is worth filing. found:false means we do not hold that CARC in our dataset (X12 publishes far more codes than we have curated) - the response still returns the group-code meaning when a group prefix (CO/PR/OA/PI) was supplied, and note explains the gap rather than leaving it silent. unknownRarc reports, rather than silently drops, any RARC you passed that we do not hold. Passing payer additionally attaches that payer's appeal deadline where we hold one. Free, no API key needed.
    get_ncci_editsLook up NCCI procedure-to-procedure (PTP) edits, in two modes. Pair mode: pass column1, column2 (either order) and dateOfService to check whether that specific pair bundles on that date; status is no_data (we hold no edit for this pair in our dataset - NOT confirmation CMS has none), not_applicable (we hold the edit but it does not apply on this date, either outside its effective/deletion window or modifierIndicator "9" meaning the edit is deleted or never applicable), or applies. List mode: pass code alone to get every pair in the dataset involving that code, in either column, with no date filtering. modifierIndicator "0" means no modifier may ever bypass the edit; "1" means a modifier may bypass it, but only where the services were genuinely distinct. Free, no API key needed.
    get_mueLook up the medically unlikely edit (MUE) unit limit for a single CPT/HCPCS code. status is found or no_data; no_data means we hold no MUE value for this code in our dataset, not that CMS publishes none. When found, the returned limit.mai (adjudication indicator: "1" line edit - denies the excess units on the line; "2" date-of-service absolute - never payable above the limit on that date, not even on appeal; "3" date-of-service clinical - may be allowed above the limit with supporting documentation) governs what a biller can do above the limit, explained in limit.maiMeaning. Free, no API key needed.
    draft_appealGenerate an appeal letter for a denial, assembled deterministically from cited facts. REQUIRES A PRO OR SCALE PLAN. Called by an anonymous caller or a key on the free plan, this returns an upgrade_required error naming the pricing page - it never fabricates or partially generates a letter for a caller who cannot access the feature. The letter argues the billing question only: it never asserts anything about the patient's clinical condition. Any fact only the practice holds (providerName, claimNumber, dateOfService, codes, the signature) that is not supplied is rendered as an explicit "[TO BE COMPLETED BY PRACTICE]" placeholder in the letter body and listed by name in placeholders, never invented. grounded is true only when we hold the CARC supplied and could argue it with our own corrective-action data; when false, the letter still assembles around payer/claim details and any practiceNote given, but the substantive grounds section is left as a placeholder for the practice to write.
    get_timely_filingDays remaining to file an initial claim (or, if already past, how far over) for a payer given a date of service. daysRemaining and expired are null when we hold no published timely filing limit for that payer, which is common since many limits are contract-specific - check note before treating the result as an actual deadline. Free, no API key needed.