wealthmcp.ai
← Back to directory
MCP-Native AppOfficial

Read-only DOL Form 5500 research with nine MCP tools and nine instruction-based skills for plan diligence, fee transparency, peer benchmarking, compliance disclosures, committee review inputs and prospecting. Search Schedule C by service-provider name or EIN, combine provider and plan filters, and retrieve filed compensation alongside derived provider plan counts. Includes filed plan profiles, filing-history comparisons, cohort statistics and charts, with no account or API key required.

form 5500retirement plansservice-provider feesprovider searchSchedule Cplan prospectingcohort percentilesDOLmcp-native
Status
Official
Auth
None
Hosting
Remote
Community rating
Not rated yet
Role
Domain

Specialized public intelligence. Turns a public dataset into a queryable, agent-ready domain expert. More on Domain servers → All seven roles →

Open documentation →

Caveats

  • Reviewed the October 23, 2026 v1.0.1 release. Independently checked public tool discovery and a provider-name search without credentials: nine tools were exposed, the search returned filed Schedule C compensation and derived provider identities, and server_info reported server 1.1.0 and nine skills. Software release labels and the live server version differ. Filing accuracy, complete provider coverage, cohort calculations and chart output were not independently audited. Sources: https://5500mcp.ai/release-notes and https://5500mcp.ai/methodology.
  • Coverage is the loaded filing population, not every U.S. plan or live plan operations. On October 23, server_info reported 164,657 plans, 390,591 filings, filing years 2024 and 2025, and a September 3, 2026 dataset refresh date; the sampled search row reported September 2. Check result-level vintage and filing year rather than assuming daily freshness. DOL datasets refresh monthly and filings can be late or amended. Confirm figures against the plan's own EFAST2 filing.
  • Provider search covers Schedule C filers only, mainly plans with 100 or more participants; 5500-SF filers and providers paid under $5,000 are not visible in this search, according to the vendor. Exact normalized-name matching is the default; contains matching is for discovery. An EIN-plus-name query falls back to exact names where the filed EIN is blank. Name variants, absent EINs and similarly named firms can make identity matching incomplete or ambiguous. Provider plan counts are derived over filtered matches, not market share or a verified current client list. The site's ADVmcp links are labelled name searches, not confirmed cross-dataset identities.
  • Schedule C compensation and Schedule H administrative-expense lines cover different reporting bases; a gap is not automatically an error, and filed compensation is not a complete all-in fee or provider-revenue estimate. Missing values remain not filed, never zero. Fee percentiles require the cohort definition, cohort size and separate filed-fee coverage; they are positions, not conclusions that fees are high, low or reasonable. Ratios and provider counts are derived arithmetic, not filed figures.
  • Plan-change comparisons distinguish filed-value changes, first capture and same-year restatements; they do not establish cash flows or explain a change. Schedule H Part IV answers are self-reported disclosures, not findings of misconduct. Committee briefs and prospecting screens supply research inputs, not fiduciary, investment or legal advice, endorsements, compliance determinations or recommendations to approach a plan.
  • Public MCP access is account-free but subject to fair-use limits, not anonymous or zero-retention. The published privacy policy describes temporary IP-based rate-limit counters and pseudonymous usage metadata, including hashes of request arguments, retained for 180 days; it says raw argument values are not written to the usage-log table. Infrastructure still receives standard HTTP request data. Source: https://5500mcp.ai/privacy.
  • Refresh or reconnect older clients to load the new optional provider-search parameters. Skills are instructions fetched at call time, not nine additional MCP tools; they update without reconnecting. Tool-list change notifications help event-capable clients but do not guarantee every client's cached schemas refresh.
  • Start with read-only scopes and keep client-side approval prompts on until the connection is proven.
Rate this server

How to connect 5500mcp

Step 1

Remote server: paste a URL

5500mcp publishes a remote (Streamable HTTP) endpoint, so nothing is installed locally. Add it as a custom connector in Claude, ChatGPT or your IDE and paste the endpoint below.

Endpoint

https://5500mcp.ai/mcp

Client config snippet

{
  "mcpServers": {
    "5500mcp": {
      "url": "https://5500mcp.ai/mcp"
    }
  }
}

Step 2

Authenticate with no credentials

No credentials are required for public data. Rate limits still apply, so avoid bulk pulls from a chat client.

Step 3

Verify, then widen access

Restart your client, confirm the server's tools appear, and run one read-only question first. Keep client-side approval prompts on for any tool that can write, trade or send. Confirm with your firm's compliance policy before connecting systems holding client data.

Prompts that use this server

Starter prompts from the Practice Stacks that rely on Retirement & Form 5500 Research.

Retirement income sequencing draft

A draft withdrawal order with the reasoning shown.

Financial PlanningFinancial PlanningRetirement & Form 5500 Research

Read-only: draft a retirement income sequencing outline for <client name> across taxable, tax-deferred and Roth accounts, including RMD timing and a rough tax-bracket view. Pull household balances and goals from the planning tool. Use Form 5500 data only for what it can prove: identify each employer plan and sponsor, confirm the recordkeeper/provider and auditor, read participant counts and total plan assets, and check plan characteristic codes for features that are actually disclosed there (for example: Roth deferrals, participant loans, hardship withdrawals, auto-enrollment, DB vs DC). Do not infer distribution rules, in-service withdrawal availability, QMAC/QNEC treatment, or other plan-specific terms from the Form 5500. For every workplace account rule used in the sequence, require the source to be the Summary Plan Description or plan document. Not the 5500 filing. Show your reasoning, every assumption, the source for each account's rules, and flag any rule that still needs the SPD or plan-document confirmation. This is a discussion draft, do not update the plan, and note where a CPA should confirm the tax treatment.

Guardrail · Draft only. Tax conclusions need a qualified professional's confirmation.

Plan profile from Form 5500

A factual snapshot of a plan before the first conversation.

Retirement Plan Due DiligenceRetirement & Form 5500 ResearchCompliance, Risk & Regulatory

Read-only: pull the most recent Form 5500 data for <plan sponsor name> and summarise participants, assets, provider, recordkeeper, auditor and any reported issues, then check the named provider and advisor against my compliance and regulatory research tools for registration status and disclosure history. Include the filing year, cite each figure, and show filing facts and regulatory findings in separate columns. Mark anything not in the filing or the regulatory record as 'not disclosed' rather than estimating.

Guardrail · Read-only public filings. Confirm currency of the data before using it.

Plan fee benchmarking read

How the plan's costs compare, with sources shown.

Retirement Plan Due DiligenceRetirement & Form 5500 ResearchCompliance, Risk & Regulatory

Read-only: from the Form 5500 and schedules for <plan name>, summarise disclosed fees per participant and as a share of assets and compare to plans of similar size using only sources you can cite, then note any fee-related disclosure, complaint or exam finding for the plan's provider or advisor from my compliance and regulatory tools. State the limitations of both the benchmark and the regulatory record. No recommendations to switch providers, just the analysis with each figure's source.

Guardrail · Read-only. Benchmark data varies in quality. Always show the source.

Read the full connection guide →