Investor overview · 2026

Beleggo

The provably-local engine for Europe's e-invoicing mandates — validate, fix and generate compliant invoices without a single byte leaving the device.

Beleggo · beleggo.de · Seed

01

Why now

Europe is being legally forced onto structured e-invoicing.

The legislator sets the demand, on a fixed schedule, country by country. Nobody can opt out.

The demand clock is already ticking

🇩🇪
2025
Germany — receive obligation, every B2B
🇫🇷
2026-09
France begins issuing · BE / PL / GR live
🇩🇪
2027
Germany issue — turnover > €800k
🇩🇪
2028
Germany issue — everyone
🇪🇺
2030
ViDA — intra-EU B2B, EN 16931

Two German purchase-urgency waves and France are ahead of us, not behind. Every euro invested in the EN 16931 core keeps compounding toward 2030.

Total addressable market

€8.9B→€23.7B

The EU e-invoicing & compliance market, 2024 → 2028 — compounding at ~28% a year.

Source: Billentis "Riding the Tornado", June 2026. A rounding-error share of this is a real business.

02

The problem

Structured invoices are brittle — and confidential.

Two problems in one file. Both are getting worse as volume mandates land.

One malformed field → silent rejection

An XRechnung ZUGFeRD Peppol file carries ~1,700 rules. Miss one and the invoice is silently rejected by the buyer's system — or the tax authority's.

The rules change on legally-mandated dates, several times a year. A validator with stale rules gives wrong compliance verdicts — the product silently failing.

The other half of the problem

The invoice carries customer names, bank details, revenue.

65%
of DACH firms name data-protection as the #1 blocker to adopting tooling

The dirty secret of the "local" tools

The obvious fix is a desktop validator. So we opened the leading paid one and read its own docs.

"The software utilizes an external validation service, operated on a secure server by 7-PDF."

The paid "desktop" competitor uploads every invoice. A genuinely on-device validator is far rarer than the market looks — and unlike a claim, this is checkable: open DevTools, pull the network cable, it still works.

03

The solution

Beleggo runs the whole compliance engine on the device.

Validate · auto-fix · generate · convert — browser, desktop, CLI, or AI agent. Nothing uploaded, ever. Verifiable by anyone.

What it does today

VALIDATE

5 layers deep

XSD → EN 16931 Schematron → PDF/A-3 container → page↔XML reconciliation → B2G preflight. 1,749 rules.

FIX

Invalid → valid

One-click safe repairs, a guided editor, and a normalize tier — without changing a monetary value.

GENERATE

Create & convert

UBL ↔ CII ↔ ZUGFeRD/Factur-X PDF, validated before download.

FORMATS

8, one engine

XRechnung · Peppol BIS 3 · NL / RO / HR / PT CIUS · EN 16931 core.

SURFACES

Everywhere you work

Offline web PWA · desktop app (mac/win/linux) · CLI · a local MCP server.

LANGUAGES

DE · EN · FR

Invoices generate in the seller's language; the binding XML is identical.

It's real, and it's proven

$ beleggo validate rechnung.xml
  ✗ invalid · XRechnung UBL · EN 16931
  BR-DE-15  Buyer reference (BT-10) missing
  → auto-fixable

$ beleggo fix rechnung.xml --validate
  invalid → valid
  wrote rechnung.fixed.xml

Cross-checked against the officials

Verdicts asserted against the KoSIT jar, Service-BW, the EU ITB and VeraPDF — the same engines behind the government routers.

13/13
conformance cases agree
208
sample data kit · 0 disagreements

04

Defensibility

Three moats you can't decompile.

The rules are public and the engine can be rebuilt. These can't be copied.

Why we win

01

Provably local

The one claim server-side incumbents structurally cannot make — aligned with the #1 stated market blocker, and checkable by anyone.

02

The freshness treadmill

Rules change on legal dates. We ship the updates so the customer doesn't watch six national feeds. Recurring by construction.

03

The referee

The 208-sample data kit + cross-validator report = we sell the proof your software is right. Vendor-neutral trust compounds.

Business model — a ladder, not a price

Free
€0
Studio · CLI · MCP. The funnel & agent-discovery channel.
Pro · pay-once
€99
SMB & Kleinunternehmer. Incl. 12 mo of rule updates.
Kanzlei / kit
€390–1.5k/yr
Tax practices, QA & data-kit teams.
OEM / white-label
€2.9–15k/yr
Vendors & partners who embed or resell.

Everything above €99 includes updates only while subscribed — the artefactsUntil license claim turns pay-once into recurring, on a legislator-driven renewal cycle.

Distribution — the channel incumbents can't follow

The agent channel

MCP SDK usage: 2M → 97M / month in a year. Beleggo is a local MCP server an AI agent calls to validate & fix invoices inside the user's workflow.

A surface Google can't disintermediate and server-side rivals can't enter without breaking their own architecture. Classic programmatic-SEO is dying — Snyk's flagship SEO tool was shut down Jan 2026 as LLMs ate search.

White-label & OEM

Client-side architecture = each partner deployment is a static bundle — near-zero to operate, no per-tenant data liability.

A partner is the sales channel a small team can't build: Kammern, Steuerberater networks, ERP vendors. Server-side competitors can't white-label without hosting the data.

Competition

 
Free Java
Hosted SaaS
7-PDF
Beleggo
Provably local (no upload)
server-side
uploads
uploads
✓
Product surface · UI/CLI/MCP
library
✓
thin client
✓
Auto-fix · container · page↔XML
✗
partial
✗
✓
Runs in the browser / on device
JVM
✗
✗
✓ WASM
Multi-country · one engine
manual
✓
DE only
✓ ×8

Mustang / phive-rules are excellent and free — but Java libraries for integrators, not a product (43 stars despite 40+ formats). Distribution is the bottleneck we solve.

Let the AI write the invoice — locally, we're the one who checks it.

The deterministic guardrail for a probabilistic AI-accounting wave — a position the incumbents cannot copy.

The ask

We've built the engine, the proof and the channel.

Fund the go-to-market to own the provably-local layer of EU e-invoicing — while the legislature keeps scheduling demand for 2026-09, 2027, 2028 and 2030.

Beleggo · [email protected] · beleggo.de