Skip to main content
Introducing MyVoteCounts: Maryland-first civic infrastructure, built for clarity and public trust.
BallotBox, Inc. Strategy. Technology. Trust.

AI & Civic Data Infrastructure

Verified civic context for the AI era.

BallotBox is building a reusable civic intelligence layer that transforms fragmented official information into normalized, source-visible context for public products, professional workspaces, developers, institutions, and AI systems.

Strategic initiative

Platform development

Design and specification work centers on the canonical civic core, State Packs, Civic Assurance, and future MCP and API distribution.

Core doctrine

Models reason. Tools verify. Sources establish civic truth.

BallotBox is designing more than an AI layer on top of static data. The infrastructure is built to adapt across models, data sources, jurisdictions, products, workflows, election cycles, and governance requirements.

AI can interpret questions, organize research, explain context, and help systems reason. Deterministic tools resolve civic geography and structured facts. Official sources, provenance, freshness, corrections, and human accountability establish the evidence behind every civic answer.

System architecture

One civic intelligence system, distributed across every experience.

BallotBox civic intelligence system architecture. Seven layers in order: Official Source Network, Source Registry, Canonical Civic Core, State and Local Adapters, Civic Assurance, Distribution Surfaces, and Product Experiences. Downstream examples include MyVoteCounts, BallotBox Pro, AI platforms, developers, civic institutions, researchers, and media.
  1. Official Source Network

    Election authorities, legislative records, geographic files, public offices, calendars, rules, results, and other authoritative civic sources.

  2. Source Registry

    A structured inventory of source ownership, authority, coverage, effective dates, retrieval dates, versions, rights, and operational status.

  3. Canonical Civic Core

    A shared schema for addresses, jurisdictions, districts, offices, officials, elections, voting methods, source evidence, and civic relationships.

  4. State and Local Adapters

    Jurisdiction-specific mappings that translate each state’s geography, election administration, terminology, rules, and source systems into the canonical core.

  5. Civic Assurance

    The provenance, freshness, correction, evaluation, monitoring, versioning, and review system that supports every distribution channel.

  6. Distribution Surfaces

    MCP, API, web experiences, civic packets, developer documentation, structured pages, and professional product integrations.

  7. Product Experiences

    MyVoteCounts, BallotBox Pro, AI platforms, developers, civic institutions, researchers, media, and public-sector users.

Downstream examples

  • MyVoteCounts
  • BallotBox Pro
  • AI platforms
  • Developers
  • Civic institutions
  • Researchers and media

Infrastructure family

Four infrastructure products. One assurance-backed system.

  • Civic Context MCP

    Strategic initiative

    A planned Model Context Protocol interface designed to ground AI hosts and agents with structured, source-visible civic context.

  • Civic Context API

    Strategic initiative

    A planned developer and institutional interface for normalized civic data, provenance, evidence status, freshness, and correction history.

  • State Packs

    Planned platform

    Planned jurisdictional operating packages that connect one national civic core to the sources, geography, rules, calendars, and operating realities of every state.

  • Civic Assurance

    Planned platform

    The planned quality, provenance, monitoring, evaluation, and correction layer supporting every product and distribution surface.

Civic Data Infrastructure

Strategic initiative

Civic Context MCP

Verified civic context for every AI system that needs to get democracy right.

A proposed read-only Model Context Protocol server for source-cited, versioned, provenance-first civic context.

For: AI platforms · Developers · Civic organizations · Universities · Media · Public institutions

Infrastructure principles

  • Read-only public civic context
  • Deterministic civic tool design
  • Source-visible responses
  • Public and professional infrastructure separation

Infrastructure development

Civic Context MCP is a strategic initiative in design and specification. Current work centers on tool definitions, resource schemas, source policy, evaluation requirements, and future public distribution.

Civic Data Infrastructure

Strategic initiative

Civic Context API

One trusted interface to fragmented American civic data.

A proposed developer and enterprise interface for normalized civic context, sources, evidence labels, effective dates, and correction history.

Infrastructure principles

  • Versioned civic contracts
  • Inspectable source evidence
  • Developer and institutional integration
  • Shared canonical civic core

Infrastructure development

Civic Context API is a strategic initiative in architecture and specification. Current work centers on versioned contracts, canonical schemas, provenance-rich responses, and developer integration patterns.

Civic Data Infrastructure

Planned platform

State Packs

Fifty states. One national core. State-specific intelligence.

State Packs are the planned operating packages for state identity, election authorities, geography, offices, voting rules, calendars, sources, GIS, provenance, monitoring, and launch controls.

Infrastructure principles

  • Canonical national core
  • State and local adapters
  • Versioned readiness manifests
  • State-specific source operations

Infrastructure development

State Packs are planned infrastructure. Maryland serves as the first deep-validation environment for a national system built from one canonical core and 51 jurisdictional operating packages covering the 50 states and the District of Columbia.

Civic Data Infrastructure

Planned platform

Civic Assurance

The assurance system is the product.

The planned assurance layer combines a source registry, civic ontology, normalization, versioned provenance, correction history, evaluation sets, freshness monitoring, and controlled publication.

Infrastructure principles

  • Provenance across every layer
  • Freshness and correction operations
  • Evaluation before release
  • Versioned accountability

Infrastructure development

Civic Assurance is planned infrastructure. Current work defines the source registry, provenance model, freshness operations, correction lifecycle, evaluation sets, version controls, and publication architecture.

Platform foundation

A civic intelligence core designed for change.

  • Canonical Civic Schema

    Create stable shared definitions for civic geography, offices, elections, officials, voting methods, sources, evidence, and relationships.

  • Source Registry

    Maintain a structured operating record of authoritative sources, ownership, coverage, versions, effective dates, retrieval dates, and status.

  • Civic Ontology

    Define how civic entities relate across federal, state, county, municipal, legislative, election, and administrative systems.

  • Deterministic Tools

    Resolve civic geography, offices, representatives, calendars, voting methods, and other structured facts through reviewable tool paths.

  • Versioned Manifests

    Preserve data, source, schema, tool, adapter, and release versions so civic answers can be inspected and reproduced.

  • Freshness Operations

    Monitor expected update cycles, aging sources, election-cycle changes, jurisdictional revisions, and operational review priorities.

  • Correction Lifecycle

    Record corrections, supersession, propagation, review, and release across products, caches, packets, APIs, MCP resources, and evaluations.

  • Evaluation and Monitoring

    Use gold evaluation sets, regression tests, tool checks, source monitoring, and review workflows to measure infrastructure quality over time.

National architecture

Fifty states and the District of Columbia. One national civic core.

Maryland provides the first deep-validation environment for the BallotBox civic architecture. National scale comes from preserving one canonical core while State Packs and local adapters carry the geographic, legal, administrative, source, and election-cycle differences of each jurisdiction.

The result is a reusable system that can expand state by state while maintaining shared contracts, common evidence standards, comparable operating controls, and a visible coverage record.

AI discovery

Build civic infrastructure that machines can find, inspect, and cite.

BallotBox is designing the public information layer for both traditional search and AI-mediated discovery. Authoritative product pages, canonical URLs, structured metadata, source and vintage blocks, citations, documentation, changelogs, API specifications, MCP resources, and coverage manifests create a stronger foundation for machine retrieval and civic citation.

Build the civic context layer with us.

BallotBox is preparing technical briefings and design conversations with AI platforms, developers, civic organizations, universities, media organizations, public institutions, election specialists, and infrastructure partners.