Whitepaper v1.0
Entity Identity Layer (EIL)
Infrastructure for the Agent Web
Verified, machine-readable identity for organizations and persons — domain-bound, agent-native, canonical.
Version 1.0.0 · EIL Card / Sinyalle pilot · June 2026
Executive summary
Today's web was built for human eyes: pages, layouts, and funnels. The agent era pushes autonomous systems to infer who an entity is from megabytes of HTML, SEO snippets, and search noise — wasting compute, amplifying hallucination risk, and weakening trust boundaries. Entity Identity Layer (EIL) introduces a domain-bound, machine-readable canonical record that agents read via resolve() in milliseconds. EIL is not a payment card or user login platform; it is identity infrastructure for the agent web.
1. Problem statement
Three structural limits in the human-centric web slow autonomous AI adoption:
Compute and energy waste
Agents download, strip, and tokenize HTML, CSS, and JavaScript to learn an official name or product list — work that should be a single JSON read.
Identity ambiguity
TLS secures the pipe, not the meaning behind the domain. Without a verified entity layer, models guess from snippets and training noise.
Structured data mismatch
Schema.org helps search engines index pages for humans. It is not optimized as the primary trust boundary for agent function calling and registry-backed resolve().
2. The solution: EIL
EIL lets organizations publish one authoritative Digital Card: canonical JSON tied to a domain, exposed through a public registry and optional /.well-known/digital-card on the entity's own origin. Agents call DigitalCard.resolve({ domain }) or hit the registry API directly. Verified records expose verified: true after DNS (or other) attestation — a machine trust signal distinct from transport security.
3. Technical architecture & discovery
EIL uses a hybrid discovery model: registry for speed and indexing; well-known for domain-bound proof; DNS TXT for ownership verification. The TypeScript SDK tries registry first, then falls back to the domain well-known URL.
Hybrid discovery flow (SDK resolve)
[AI Agent / SDK]
│
├── 1. Registry (primary)
│ GET /api/v1/resolve?domain={domain}
│
├── 2. Well-known (fallback + domain proof)
│ GET https://{domain}/.well-known/digital-card
│
└── 3. DNS TXT (verification signal)
_digital-card.{domain}4. Data schema specification
EIL Card schema v1.0 discriminates organization and person profiles. The example below matches the live registry format — not a fictional alternate schema. Full reference: schema v1.0 and @digitalcard/sdk.
Organization card (illustrative — schema v1.0)
{
"schema_version": "1.0",
"card_id": "example.com",
"type": "organization",
"handle": "example",
"verified": true,
"verification_method": ["dns"],
"name": {
"official": "Example Technology Inc.",
"short": "Example"
},
"contact": {
"email": "info@example.com",
"website": "https://example.com"
},
"description": {
"tagline": "Infrastructure for autonomous agents",
"summary": "Canonical organization profile for AI agents — no HTML scraping required."
},
"products": [
{
"id": "agent-core",
"name": "Agent Core API",
"description": "High-throughput machine routing interface."
}
],
"same_as": ["https://github.com/example"],
"updated_at": "2026-06-17T00:00:00.000Z",
"human_url": "https://eilcard.com/kart/example",
"registry_url": "https://eilcard.com/api/v1/cards/example"
}5. Security & trust framework
Trust is layered: transport (TLS), attestation (DNS verification), and canonical source (registry + synced well-known).
DNS domain verification
LiveDashboard TXT challenge at _digital-card.{domain}. Verified cards expose verified: true and verification_method: ["dns"].
Registry as authoritative source
LiveRegistry updates first; domain well-known can proxy or mirror registry JSON (nginx proxy pattern). SDK compares updated_at when both surfaces exist.
JSON Web Signatures (JWS)
RoadmapSigned card payloads for tamper detection on self-hosted well-known — planned for schema v1.1.
DNSSEC guidance
RoadmapEnterprise hardening for TXT record integrity — documentation and optional verification, not required for MVP.
6. Ecosystem integration & roadmap
Adoption surface shipped or planned:
Registry + SDK + well-known
Live@digitalcard/sdk, eilcard.com registry API, dashboard discovery panel, agent-card.json template, llms.txt exports.
Agent integration docs
Live/docs/agents — system prompts and function-calling templates for OpenAI, Anthropic, and Gemini.
Framework tools
RoadmapLangChain EILResolveTool, LlamaIndex reader, OpenAPI spec, MCP resolve tools.
Schema.org bridge
RoadmapOptional resolve fallback that infers organization JSON from page JSON-LD when no EIL card exists — meta.source: schema.org-inferred, verified: false.
Standards track
RoadmapIANA well-known URI registration and Internet-Draft process after spec stabilization.
Live pilot
Sinyal 24 (sinyalle.com) — resolve, well-known proxy, and agent discovery exports.
eilcard.com/kart/sinyal24 →7. Conclusion
The web is gaining a machine-facing layer. Agents should not parse homepages to learn who stands behind a domain. EIL delivers verified, canonical entity JSON — the identity infrastructure for autonomous systems. The pilot is live at eilcard.com with domain-bound well-known on participating origins.