TaxIntelligence

0
IRC Subtitle A — fully indexed & queryable via REST

The IRC Tax Code
as a Knowledge
Graph

1,133 TaxSection nodes. 5,478 relationship edges. Every cross-reference, exception, and dependency mapped. Natural language in, Cypher graph traversal out. Queryable via REST API.

1,133

TaxSection Nodes

5,478

Relationship Edges

IRC

Subtitle A Coverage

REST

API Access

The Problem

4 million words.
Zero relationship
context.

The IRC is the most cross-referenced legal document in America. Current tools treat it as flat text. They search keywords. They can't traverse cross-references, detect exceptions, or map dependency chains. Every missed link is a missed deduction — or a compliance risk.

Keyword search is blind

Searching "199A" returns the section. It doesn't show you that 199A(d) references 469, 170, and a dozen other sections that modify the deduction. You get a page, not a map.

Cross-references are invisible

Section 162 references 274, 280A, and 195. Section 280A references 162 back. These circular dependencies create traps that flat-text tools can't surface.

Exceptions hide in plain sight

"Except as provided in subsection (d)" — buried 3 levels deep. A human might miss it. A graph traversal finds every exception path automatically.

Graph Traversal 1,133 Sections 5,478 Edges REST API Zero Copyright Risk Graph Traversal 1,133 Sections 5,478 Edges REST API Zero Copyright Risk

Capabilities

Graph-powered
tax intelligence

Graph Traversal, Not Keyword Search

Every query traverses the actual relationship graph. Ask about Section 199A and get the full dependency tree — cross-references, modifications, exceptions, and downstream effects — not a list of keyword matches.

BFS

Breadth-first

DFS

Depth-first

SP

Shortest path

Cross-Reference Mapping

5,478 edges map every "as defined in," "except as provided," and "subject to" across the IRC. See how sections depend on, modify, and override each other.

Exception Detection

Automatically surfaces exception paths and carve-outs. Never miss the "except as provided in subsection (d)" buried three levels deep.

Natural Language Queries

Ask in plain English. The API translates your question into a Cypher graph traversal and returns structured results with relationship context.

REST API Access

Authenticated REST endpoints on port 8084. X-API-Key auth. Search, document retrieval, domain taxonomy, team feeds, and metrics — all JSON.

Public Domain Data

Sourced exclusively from GovInfo USLM XML and eCFR — public domain, zero copyright risk. Redistribute, embed, or resell without licensing concerns.

Real-Time Updates

Graph stays current with IRC amendments. New sections and relationship changes are ingested as GovInfo publishes updates. No stale data.

Bulk Export

Export entire subgraphs, domain taxonomies, or filtered node sets as JSON. Build your own analytics, feed internal tools, or populate your own database with structured IRC graph data. Enterprise plans include full graph snapshots.

JSON CSV GraphML Export formats

Live API

Query the graph

One REST call. Natural language in, structured graph data out. Every response includes the section, its title, cross-references, and edge count.

Request

# Search for a section by natural language

GET /v1/search?q=section 199A qualified business income deduction

# Headers

X-API-Key: your-api-key

Content-Type: application/json

Response — 200 OK — 47ms

{

  "results": [

    {

      "section": "199A",

      "title": "Qualified Business Income",

      "subtitle": "A",

      "chapter": "1",

      "cross_refs": ["170", "469", "199A(d)", "1(h)(11)", "751(a)"],

      "edges": 23,

      "relationship_types": ["REFERENCES", "MODIFIES", "EXCEPTS"]

    }

  ],

  "traversal_depth": 2,

  "query_time_ms": 47

}

/v1/search

Natural language search

/v1/sections/:id

Document retrieval

/v1/taxonomy

Domain taxonomy

/v1/metrics

Graph metrics

Pipeline

How it works

Step 01

Ingest Public Law

GovInfo USLM XML and eCFR data is parsed into structured nodes. Every section, subsection, and paragraph becomes a typed node in the graph. Public domain — zero copyright risk.

Source

govinfo.gov/uslm/title-26

ecfr.gov/title-26

Public domain. No license required.

Step 02

Build the Graph

NLP extraction identifies every cross-reference, exception clause, modification, and dependency. Each becomes a typed, directed edge in Neo4j. 5,478 edges across 1,133 nodes.

Edge Types
REFERENCES (cross-citations)
MODIFIES (amendments)
EXCEPTS (carve-outs)
DEPENDS_ON (prerequisites)
DEFINES (term definitions)

Step 03

Query via API

Natural language questions are translated to Cypher graph traversals. The API returns structured JSON with the section, cross-references, edge counts, and relationship types. Sub-50ms response times.

Performance
Sub-50ms graph traversals
Structured JSON responses
X-API-Key authentication

Step 04

Integrate Everywhere

Feed graph data into your tax software, compliance platform, or research tool. Bulk export entire subgraphs. Build internal products on top of structured tax intelligence without maintaining the graph yourself.

Integrations
Tax preparation software
Compliance engines
Legal research platforms

Pricing

Graph intelligence
at scale

Explorer

$300 /month

For individual practitioners and small firms exploring graph-powered tax research.

  • 100 queries / month
  • Natural language search
  • Cross-reference mapping
  • Section document retrieval
  • JSON export
  • Email support
Get Started

Enterprise

Custom

For legal tech platforms, fintech compliance teams, and organizations building on the tax graph.

  • Unlimited queries
  • Everything in Professional
  • Full graph snapshots
  • Custom edge type queries
  • Webhook notifications
  • Dedicated account manager
  • SLA + onboarding
Contact Sales

Comparison

Not another
search tool

Feature Tax Intelligence Westlaw Bloomberg Tax
Graph Traversal
API Access Limited Limited
Cross-Reference Mapping 5,478 edges Hyperlinks only Hyperlinks only
Exception Detection
Natural Language Queries Basic Basic
Price From $300/mo $200–$600/mo $400+/mo
Copyright Public domain Proprietary Proprietary
Bulk Export

Try 10 free
queries

No credit card. No contract. Hit the API, traverse the graph, see the cross-references your current tools miss. Ten queries is enough to know if graph intelligence changes your workflow.

X-API-Key Auth Sub-50ms Response Public Domain Data