Skip to main content
Public-interest infrastructure · Pre-501(c)(3)

Funding the public proof layer for an AI-era legal system.

NORA Foundation builds open, cryptographically verifiable infrastructure so that anyone — without our cooperation — can prove a court record, exhibit, or claim is what it says it is. The same primitive serves pro se litigants, journalists, public defenders, and watchdogs.

Tens of M
Civil pro se litigants annually (Self-Represented Litigation Network — srln.org)
$0
Cost to verify a NORA Canon attestation
5
Interlocking systems on one shared protocol
12 mo
Capital ask covers v1 launch of public tools

Pro se prevalence sourced from the Self-Represented Litigation Network (srln.org). NORA does not publish a single point estimate; civil pro se appearances are tracked in the tens of millions annually across U.S. state courts.

Thesis

A neutral, cryptographic primitive for “this record is real” is the missing piece of civic AI infrastructure.

Every other public-good infrastructure layer has a verifier of last resort: TLS for the web, DNSSEC for naming, Signal for messaging, Bitcoin for value. Legal records do not.

NORA Canon is that primitive. It is small, falsifiable, and runs on commodity infrastructure. Once a record is sealed under Canon, no party — including NORA — can quietly alter it without the alteration being mathematically obvious to anyone who looks.

We are not building a court. We are building the verifier the courts, journalists, and pro se litigants can share.

Problem

What breaks when records can't be verified.

Records can be edited, suppressed, or fabricated.

Court records are the only thing that survives bad-faith institutions, yet they are routinely altered, sealed without notice, or quietly disappeared. The people who depend on them most have the least leverage to verify them.

Verification today requires institutional cooperation.

To prove a docket entry is real, you currently need the court (which may be the bad actor) or a paid service (which charges per-look). There is no neutral, public, cryptographic primitive for 'this record is what it says it is.'

AI is about to multiply the problem.

Generative AI makes synthetic court records, fake screenshots, and convincing fabricated transcripts trivial. Without a falsifiable trust layer, the default assumption about any digital legal record will become 'could be fake.' That breaks the system.

Why now

The window is open for the next 18 months.

  • The Wisconsin and Minnesota courts have public dockets but no machine-verifiable provenance — we can attest them today.

  • Cloudflare AI Gateway + Ed25519 + content-addressed storage make a public attestation lane affordable on commodity infrastructure.

  • Foundations from civic-tech to AI-safety are explicitly funding infrastructure for "AI-era authenticity."

  • Pro se litigation is increasing in every state. Cost-of-counsel is rising; the access gap is widening.

Market

A primitive that compounds across four overlapping audiences.

The protocol is the same. The wrapper changes. Each audience uses the same Canon-attested records for a different purpose.

Pro se litigants
Tens of millions annually

civil pro se appearances per the Self-Represented Litigation Network (srln.org) — family, housing, debt, small claims.

Investigative journalists
~1k orgs

need verifiable court-record snapshots; today pay PACER + private services.

Public defenders / legal aid
~6k orgs

understaffed; tools are decades behind the bar.

Watchdog NGOs
~3k orgs

oversight depends on records that the institution being watched controls.

Traction

What's shipped, what's building, what's open.

NORA Canon v0.2.0 published

Source-available specification. Open peer review. Source repo + license active.

agent@norafoundation.io live

AI email agent on Cloudflare Workers + Durable Objects + MCP server. Anyone can email it; an AI drafts a reply for human review. Case-specific or legal questions get the same refusal as NORA-1 and a referral to a state lawyer-referral service.

NORA-1 conversational guide

Live on this site. Same Claude/Cloudflare lane as agent@. Demonstrates the platform voice for funders today.

LawLlama three-plane architecture

FastAPI + Google ADK + sandboxed workspaces. Codebase active in private repos. Member access planned Q3 2026.

CASEFORUM curation pipeline

WI + MN court-records ingestion working. Initial public cases targeted for v1 launch.

AAPL pre-registration open

American Association of Pro Se Litigants. Founding-member rolls open at aapl.norafoundation.io.

Maturity
Maturity ladder

Where each NORA product sits today.

Four stages, public criteria. Modeled on the CNCF Sandbox → Incubating → Graduated ladder, adapted for an attestation stack: Spec → Pilot → Production → Audited. A product only advances when the criteria are publicly met.

  1. Stage 1
    Spec

    Written, public, reviewable. No running code required.

    Requires: Published RFC or design doc, open for comment.

  2. Stage 2
    Pilot

    Reference implementation exists; in real-world use by us.

    Requires: Public code + working deployment + at least one live user.

  3. Stage 3
    Production

    Shipping to outside users on a stable interface.

    Requires: Versioned release, license stated, support channel open.

  4. Stage 4
    Audited

    Third-party security + protocol review on record.

    Requires: Published audit report from an independent party.

Current placement
  • NORA Canon v0.2

    Spec + reference impl in Meridian-Cannon, public. Third-party audit targeted Q4 2026.

    Pilot
  • CHIPS Copilot

    Public MIT release, versioned. External adoption tracked from public-launch.

    Production
  • NORA-1

    Live conversational guide on this site. Same Cloudflare lane as agent@.

    Pilot
  • agent@

    Handles inbound at agent@norafoundation.io. Worker + Durable Objects + MCP.

    Production
  • CASEFORUM v1

    Curation pipeline working. Public v1 launch targeted Q3 2026.

    Spec
  • LawLlama web v1

    Three-plane architecture documented. Browser experience targeted Q3 2026.

    Spec
Roadmap

The 12-month build.

Q2 2026
Public marketing site + AAPL pre-registration
shipped
Q3 2026
LawLlama v1 member access · CASEFORUM v1 launch (5–10 cases)
planned
Q4 2026
Canon v1.0 freeze · third-party security audit · 501(c)(3) determination
planned
Q1 2027
CASEFORUM expansion to 50+ public-interest cases · open verifier SDK
planned

Full roadmap detail: /roadmap.

Capital ask

We are raising a 12-month operating budget to ship v1.

We are explicitly pre-launch. Donations and grants now fund the build, not maintenance of an existing product. Every dollar is accounted for in the quarterly transparency report that begins at v1 launch.

We accept individual donations, recurring giving, foundation grants, restricted grants for CASEFORUM curation, and in-kind support (legal review, security audit, infrastructure credits).

Major-gift conversations and grant applications are most useful when they start with a specific milestone — see the roadmap quarters above. We can scope a restricted grant to any of them.

Use of funds
Engineering55%

Cloud infra, AI inference, audits, security review, code review.

Legal & compliance30%

501(c)(3) filing, privacy review, content moderation policy, records retention.

Operations15%

Domain, email, document storage, payment processing fees.

No founder salary until v1 ships. No paid marketing. No travel budget. Infrastructure spend is published quarterly.
Governance & risk

What we're honest about.

  • Pre-501(c)(3). IRS determination in process. Donations are currently NOT tax-deductible — we say this in writing on every donate page.

  • Pat White (Founder, J. Patrick White) is the sole officer pending board formation.

  • Fiscal sponsorship is under consideration as an interim path; we will publish the choice before accepting major grants.

  • Quarterly transparency reports begin at v1 launch — cryptographically signed under NORA Canon.

  • All commits are public on GitHub for the open-source artifacts; private repos for in-progress code will be opened on a milestone basis.

Team

Who's building.

Founder
J. Patrick White
AI Evidence Systems Founder | Product + Automation Developer

13+ years across finance, private real estate securities, and operations (40+ rental units). Builds evidence-first AI infrastructure: immutable vaults, chain-of-custody, OCR/PDF pipelines, retrieval, citation validation, and adversarial review. USC Marshall finance + former FINRA 7/63. Hands-on with Python, TypeScript, Next.js, PostgreSQL, Docker, and LLM systems.

Advisory board · forming

We are actively forming an advisory board across civic tech, public-interest law, evidence standards, and cryptographic protocol design. Interested? Email agent@norafoundation.io.

Detail: /team.

Built in public
Built in public

The last few commits across our public repos.

Live from GitHub, refreshed every ten minutes. If a row says “view on GitHub,” the public API rate-limited us — the underlying history is still right there.

For major gifts and grant officers.

We respond to grant inquiries within two business days. For restricted grants, please include the roadmap milestone you would like to underwrite.