MuniMark  ·  Municipal Credit Intelligence

Institutional credit analysis,
automated.

MuniMark reads an Official Statement and produces a structured, source-cited credit memorandum in less than 10 minutes. Save hundreds of hours on keeping up with new issuance, odd lot hunting, and more, allowing you to focus on what really matters to clients.

See a sample analysis ↓ Request access
General obligation
Charter schools
Transportation
Utilities
More sectors in development

Why MuniMark

The same output. A fraction of the time.

A thorough muni credit memo requires reading a 200-page OS, pulling audited financials from EMMA, running scenario math, and writing it all up in investment prose. MuniMark does that in one streamlined process.

01
8 hrs → 10 min
Time per credit memo

A credit memo that takes an analyst most of a day to produce runs in about 10 minutes. The pipeline handles document parsing, financial extraction, research, and prose generation.

02
6
Specialized agents

Document intelligence, structural analysis, financial modeling, research, downside scenarios, and a dedicated QA pass — each scoped to one task.

03
$0
In proprietary data

The data layer runs on EMMA, FRED, Census ACS, BLS, FEMA, and Google News — all free public APIs. No Bloomberg terminal, no data license fees.

04
100%
Source-cited output

Every factual claim in the output has a bracketed source citation. A dedicated QA agent checks number consistency across all sections before the memo is finalized.

The pipeline

Six agents. One investment memo.

Each analytical task runs in its own agent with a defined scope. Keeping them separate prevents context bleed across a long pipeline and makes the QA pass meaningful.

Inputs
Official Statement + EMMA Filings
AGENT 01
Document Intelligence
Parses the Official Statement, extracts bond structure, security pledge, covenant terms, and flags structural anomalies before any analysis begins.
Official StatementEMMA
AGENT 02
Structural Analysis
Evaluates security pledge, flow of funds, debt service structure, and covenant headroom. Identifies structural risk flags for the analyst's review.
Bond CovenantsDebt Schedule
AGENT 03
Financial Analysis
Calculates coverage ratios, leverage metrics, and liquidity indicators. Builds the five-year trend table from audited financial statements.
Audited FinancialsEMMA Filings
AGENT 04
Research & Context
Pulls issuer demographics, economic data, news coverage, and competitive landscape from public APIs — FRED, Census ACS, BLS, FEMA, and Google News.
FREDCensus ACSBLSFEMANews RSS
AGENT 05
Downside Scenario Modeling
Stress-tests the credit under base, adverse, and severe scenarios — modeling covenant compression, coverage deterioration, and debt service risk.
Financial ModelCovenant Terms
AGENT 06
QA Verification Pass
Before final output, checks number consistency across all sections, flags unsourced claims, verifies covenant arithmetic, and surfaces any contradictions in the research.
All agent outputs
Output A
Credit Investment Memo
Full structured memo in professional investment prose — 15+ sections, fully sourced
Output B
Pitch Card
One-page summary — key metrics, top risks, buy/monitor/pass recommendation

Sample output

A real credit memo, from a real deal

Produced from the Official Statement of the City of Ladysmith, Wisconsin's 2026A GO Notes. One pipeline run, no manual editing afterward.

Issuer
City of Ladysmith, WI
Series
GO Notes 2026A
Par Amount
$3,235,000
Sector
General Obligation · WI
Recommendation
⬤ Monitor

Sample output — for demonstration purposes only. Not investment advice.

Open full memo ↗

Under the hood

Built on public infrastructure

The data layer runs entirely on free government APIs. Compute costs aside, there are no ongoing data expenses to run this.

Core stack
PythonBackend orchestration and agent logic
FastAPIREST API layer and pipeline endpoints
Claude APIPowers all seven analytical agents
PyMuPDFPDF parsing and text extraction
python-docxStructured memo output generation
ReactFrontend upload and review interface
Architecture
Multi-agent Each agent scoped to one analytical task, preventing context drift across a long pipeline
Orchestrator Central coordinator passes outputs between agents sequentially, maintaining shared document state
QA pass Dedicated verification agent checks number consistency, sourcing, and internal logic before any output
External data sources — all free, all public
EMMA / MSRB free
FRED (St. Louis Fed) free
Census ACS free
BLS free
FEMA NFHL free
Google News RSS free
SEC EDGAR free

Private beta

Request early access

MuniMark is in private beta. Leave your details and I'll follow up to walk you through it.

✦

Request received

I'll follow up shortly.