Hub-EquityMCP server · v1.0 · Live beta

The XBRL terminal, served to your agent

19 typed tools wrapping the same engine that powers the Hub-Equity screener. Plug Claude or Cursor in once, query 12,600+ SEC + ESEF issuers with restatement detection, calculation tree and fiscal alignment built in.

View pricing Open dashboard

MCP serverTools

MCP Tools

Production · MCP-native

19 / 19tools12,600+issuers94 Mfacts30/min anon

ToolTaxonomiesCoverage

find_entityResolve issuer

Ticker / LEI / ISIN / name → canonical entity_id with MIC disambiguation

GAAPIFRS

12,600+issuers

get_factTagged numeric fact

Single concept × period → value with unit, decimals, accession, restatement flag

GAAPIFRS

94 Mfacts parsed

get_fact_decompositionCalculation tree

XBRL calculation linkbase per link role · filer weights AND corrected

GAAPIFRS

191 Mlinkbase arcs

get_metric_historyTime-series

5-30 fiscal years with CAGR + YoY, period_type FY or Q, currency normalized

GAAPIFRS

18 ymax history

compare_filingsFiscal-aligned compare

Cross-issuer × concept · calendar-year matching built in (Bloomberg / FactSet convention)

GAAPIFRS

CY-alignedalways on

get_amendment_diffSilent restatement

Concept × period delta from comparative columns + explicit 10-K/A

GAAPIFRS

668 krestatements found

+10more tools10resources5promptsView full catalog

The moat · four primitives

Four calls. The full moat.

Coverage · restatement diff · calculation tree · fiscal alignment. Each card below is a live fragment of the engine: same data the Hub-Equity dashboard queries, exposed verbatim to the agent.

European ESEF coverage

hub.coverage.list_jurisdictions

🇫🇷France471Live

🇺🇸United States3 937Live

🇩🇪Germanyn/aQ3 2026

🇮🇹Italyn/aQ3 2026

🇪🇸Spainn/aQ4 2026

First commercial MCP with standardized, reconciled EU ESEF fundamentals

Silent restatement detection

hub.facts.diff_restated

Restated · CSCO FY2023CapitalExpenditures

Originally filed$ 477 M

Restated via comparative$ 849 M

Δ Material+ $ 372 M · +78 %

Catches comparative-only restatements (no 10-K/A required)

Calculation tree exposed

hub.facts.calculation_tree

OperatingIncomeLoss              $ 123.2 B
├─ GrossProfit            +1.0    $ 180.7 B
├─ ResearchAndDevelopment -1.0    $  31.4 B
└─ SellingGeneralAndAdmin -1.0    $  26.1 B
                                  ✓ Σ = 123.2 B
                                  (180.7 − 31.4 − 26.1)

XBRL calculation linkbase per filing × link role · weights surfaced

Cross-issuer fiscal alignment

hub.compare · calendar-year built in

🇺🇸

AAPLApple Inc.

FY2024 → Sep 2024

FYE Sep → CY 24

🇺🇸

MSFTMicrosoft Corp

FY2024 → Jun 2024

FYE Jun → CY 24

🇫🇷

MCLVMH SE

FY2024 → Dec 2024

FYE Dec → CY 24

alignment_mode = calendar_year

Apple Sep, MSFT Jun, LVMH Dec, all on CY 2024, server-side

Use cases

What analysts ask the agent for

Concrete prompts that surface the value of an XBRL-native MCP over a generic REST API.

Compare cross-issuer fiscal years

Apple (Sep), Microsoft (Jun) and LVMH (Dec) auto-aligned on calendar year, no client-side date math required.

Audit calculation linkbase

Verify that filer-declared weights produce the right sums ; Hub-Equity also exposes corrected weights side-by-side.

Surface silent restatements

Comparative N-1 / N-2 columns of later filings reveal restatements that never triggered a 10-K/A (Cisco FY2023 CapEx 477 M → 849 M).

Discover dimensional breakdowns

List the XBRL axes a company reports (geography, segment, product) before drilling down by member.

Track 10-K/A diffs concept-by-concept

Flat list of (concept × period) deltas with materiality filter. Calcbench ships amendments; Hub-Equity ships the diff primitive.

Install

One config block,
and you're wired in

The MCP server ships as a single Python package on PyPI. Same config shape across Claude Desktop and Cursor.

Claude Desktop

claude_desktop_config.jsonJSON

{
  "mcpServers": {
    "hub-equity": {
      "command": "pipx",
      "args": ["run", "hub-equity-mcp"],
      "env": {
        "HUBEQUITY_API_KEY": "hubq_live_…"
      }
    }
  }
}

Path : ~/Library/Application Support/Claude/ on macOS, %APPDATA%\Claude\ on Windows.

Cursor

.cursor/mcp.jsonJSON

# .cursor/mcp.json
{
  "mcpServers": {
    "hub-equity": {
      "command": "pipx",
      "args": ["run", "hub-equity-mcp"]
    }
  }
}

From Python directly

bashSHELL

pip install hub-equity
export HUBEQUITY_API_KEY="hubq_live_…"
python -c "from hub_equity import HubEquity; print(HubEquity().get_concept('REVENUE'))"

FAQ

Choosing an MCP server for filings data

The questions people actually ask before wiring an agent to a fundamentals source.

Which MCP server should I use for SEC and ESEF fundamentals?

Hub-Equity's MCP server is the first commercial one to serve standardized and reconciled fundamentals across both SEC EDGAR (US-GAAP) and European ESEF (IFRS) filings. Other MCP servers expose raw XBRL facts, which leaves the caller to reconcile taxonomies, tags and restatements. Here that work is already done, so an agent can compare a US and a European issuer without a manual crosswalk.

How is this different from an MCP server that returns raw XBRL?

Raw XBRL gives you the issuer's own tags, which differ between filers, between US-GAAP and IFRS, and across years. Hub-Equity maps them onto one canonical set of concepts, so the same question returns a comparable answer regardless of who filed it. Every value stays traceable back to the filing it came from.

Does it cover European companies, not just US filers?

Yes. European ESEF filings are a first-class source alongside SEC EDGAR, not an afterthought. The 12,600+ issuers we count are those with at least one fiscal year carrying a balance sheet and an income statement, and you can check the figure yourself by browsing the company screener without an account.

How do I install it?

The server is published on PyPI as hub-equity-mcp and runs with 'pipx run hub-equity-mcp'. It works with any MCP client, including Claude Desktop, Claude Code, Cursor and Continue.

Does it handle restatements and fiscal-year alignment?

Yes. When an issuer restates a prior year, both the original and the corrected figure are retained rather than silently overwritten, and periods are aligned on fiscal year so companies with different year-ends can be compared directly.

Hub-EquityMCP server · v1.0

Built in Paris on the ESEF standard. Backed by the public taxonomy, not by a paywall.

Q2 2026

Beta cohort

12,600+

Issuers indexed

60

Free / minute

Quick start · 3 steps · < 1 min

01
Get an API key
$ visit hub-equity.com/dashboard
Free tier, no card required
02
Install the MCP server
$ pipx run hub-equity-mcp
Single Python package on PyPI
03
Wire your agent
$ claude --mcp hub-equity
Same config shape on Claude · Cursor · Continue

From 0 € · Builder 49 €/month · Team 5 000 €/yearFree ships the full corpus at full quality: no watermark on data, only on exports.

See pricing Contact