← Back to directory
MCP-Native AppOfficial

Read-only SEC Form ADV intelligence for adviser research, due-diligence briefs, acquisition screening, peer benchmarking and claim verification. Nine public MCP tools search filed profiles, page through filtered firms, screen filing changes and return cohort statistics and charts, including qualified-custodian, advisory-employee and branch-office counts. Eight skills provide guided workflows and reference material. The free, signed-in ADV Analyst offers the same tools through a web conversation; the MCP connection needs no account or key.

form ADVSEC IAPDRIA intelligenceAUMdisclosuresfiling historypeer benchmarkingacquisition screeningclaim verificationADV Analystmcp-native
Status
Official
Auth
None
Hosting
Remote
Community rating
Not rated yet
Role

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

Open documentation →

Caveats

  • Independently checked on October 23, 2026: the public MCP endpoint returned nine tools without an account or API key, and server_info reported server 1.1.0 and eight skills. Reviewed the October 7 v1.0.4 and October 15 v1.0.5 releases. Filing accuracy, nationwide coverage, AI brief quality and privacy controls were not independently audited. Sources: https://advmcp.ai/release-notes and https://advmcp.ai/methodology.
  • Public filings, not live operating data. Verify each figure's filing and feed dates against SEC IAPD; ingestion frequency does not make the underlying disclosure current. The methodology page still identifies an August 12, 2026 feed vintage, so the listing makes no daily-freshness guarantee. Regulatory AUM, filed client and account counts are not revenue, profitability, enterprise value or interchangeable household counts. State-registered advisers and broker-dealers are out of scope.
  • Qualified custodians (Item 9.F) are a count, including the firm and related persons where filed, not custodian names. Advisory employees (Item 5.B(1)) include research staff and are not a licensed-representative count. Offices (Item 1.F(5)) exclude the principal office. Missing answers stay blank, not zero, and filters on these fields exclude nonfilers; cohort coverage and each base-rate denominator matter. Item 5.K(3) indicates whether a custodian holds at least 10% of separately managed account assets, not the custodian's identity.
  • Filing changes are arithmetic differences between dated snapshots, not explanations for growth, outflows, acquisitions or misconduct. Change screens flag windows longer than 45 days; missing historical vintages can widen comparisons. Peer percentiles are descriptive context, not firm-quality scores. Claim-verification gaps are questions for the firm, not findings that a claim is false. Research only, not investment advice or an endorsement; confirm disclosures in the adviser's own filing.
  • The ADV Analyst requires sign-in and saves investigations to the user's account; public MCP access remains unauthenticated. The vendor documents 30 Analyst questions per hour and 100 per day. AI briefs can err, so inspect tool results, dates and source filings. The August 17 privacy policy still describes authentication as operator-only and does not explain the newer saved Analyst sessions; review current terms and obtain retention clarity before entering confidential information: https://advmcp.ai/privacy.
  • Fair-use limits apply to public MCP access. The published privacy policy describes IP-based rate-limit counters and pseudonymous tool-usage logs retained for 180 days, not anonymous or zero-retention access. Refresh or reconnect older clients to load the October 7 optional screening parameters; skills are fetched at call time, and the October 15 Analyst release does not change the MCP tool surface.
  • Start with read-only scopes and keep client-side approval prompts on until the connection is proven.
Rate this server

How to connect ADVmcp

Step 1

Remote server: paste a URL

ADVmcp 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://advmcp.ai/mcp

Client config snippet

{
  "mcpServers": {
    "advmcp": {
      "url": "https://advmcp.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 Advisor & Firm Intelligence / Prospecting.

Value the practice with filed peer context

What a seller should know before naming a price.

Sell / M&A ReadinessAdvisor & Firm Intelligence / ProspectingCRMAccounting & TaxCustody & Clearing

Read-only, in three passes. First gather my practice's own figures: households, AUM, recurring vs. one-time revenue, fee mix, client age distribution and top-10 concentration from CRM and custody, and revenue, expenses, owner compensation and margin from the accounting system, recording where each figure came from and marking anything you could not source rather than estimating it. Second, run HnyComb's Value my practice skill on those inputs, keeping the source note attached to every figure, and show the seven-method blend, the range and the data confidence behind it; then run its Reconcile the Numbers skill to check the figures against filed data and published medians and name each gap. Third, pull the filed peer distribution for my state and AUM band from ADVmcp and use HnyComb's Benchmark Against Cohort skill to place my result against firms like mine. Output the valuation range, the inputs table with sources, the reconciliation gaps, and where the practice sits against filed peers. No client names or account numbers anywhere in the output, and do not write, export or share anything.

Guardrail · Read-only, no client identities in the output. The valuation is a modeled estimate, not an appraisal, and peer percentiles are positions rather than grades: Form ADV carries no revenue, EBITDA or margin figures, so peer data contextualises the number and never moves it.

Screen targets and set a walk-away ceiling

What a buyer can know before a conversation.

Sell / M&A ReadinessAdvisor & Firm Intelligence / Prospecting

Read-only. Start by writing my acquisition thesis back to me in one paragraph, geography, AUM band, client mix, custodian, succession situation, and ask about anything I left ambiguous before you search. Then run HnyComb's Find and screen acquisition targets skill: ADVmcp finds candidate advisers from filed Form ADV data and HnyComb puts a low-confidence walk-away number on each, returning a ranked shortlist with the filed facts behind every match. For the one target I name, run the Buy-Side Ceiling skill to produce a defensible ceiling, using target-supplied figures where I have them and labelling every private figure I do not have as a named assumption with the effect it has on the ceiling; if I name two targets, use Compare Targets to read them side by side without picking a winner. State the confidence level and what would raise it. Do not contact, email or approach anyone, and do not draft outreach.

Guardrail · Nothing in a public filing indicates a firm wants to sell. A shortlist is research, never a recommendation to approach or offer. Private figures are either target-supplied or a named assumption, and confidence on public data alone will be low. That is the honest result, not a defect.

Compliance history and diligence file

The regulatory record a buyer's counsel will request.

Sell / M&A ReadinessCompliance, Risk & RegulatoryAdvisor & Firm Intelligence / Prospecting

Read-only: assemble my firm's regulatory record from two directions, my internal compliance or GRC systems for what they actually hold (policy versions, control testing, vendor diligence, open findings and remediation status) and the public regulatory and firm-intelligence sources for registration history, ADV disclosures and complaints. Show them side by side and flag anything disclosed publicly with no internal record, or internally tracked that the public record does not reflect. List exam findings and correspondence separately as 'to gather offline' unless a connected system genuinely holds them. Cite each source and mark anything unverified as 'to confirm'. Do not file, submit or amend anything.

Guardrail · Read-only. The internal half of this record usually lives outside any connected tool, and filings are amended through official channels only.

Read the full connection guide →

Compared with other servers