Pre-seed · London · 2026

IDENTITY.FREE

Verifiable chain-of-custody infrastructure for AI agents in regulated finance and commerce. identity.free binds humans, organisations, products, and AI agents into one trust graph — so every autonomous action carries scoped authority, revocation, and audit evidence.

AI-Native Compliance Reusable Identity Audit Evidence Fintech Guardrails

Read the ask → See the four passports → Deck

Problem · Agentic finance has no accountable edge

Agents act. Accountability breaks.

AI agents already commit code, buy goods, update records, and move money. Regulated buyers still cannot prove who authorized which agent, under what scope, for which organisation or product — and whether that authority was still valid.

01 Authority gap

No scoped mandate

Current stacks show a login or a payment approval. They do not show the full authority chain behind an autonomous action: principal, agent, policy, scope, expiry, revocation.

02 Audit gap

No reliable replay

Banks and fintechs need reconstructable evidence months later — which human, which agent, what policy, what scope, and whether the mandate had been revoked. Today that replay does not exist.

03 Trust gap

Nodes, not edges

Identity, payment, and product-passport tools each verify one party. The missing product is the signed relationship between parties — the edge, not the vertex.

Identity is not accountability. OAuth-style authorization is single-hop, and 2025-era agent payment mandates stop at the moment of approval — none of it survives a dispute, an audit, or a regulator's question.

Thesis

Worldcoin proved humanity.
ENS proved naming.
Stripe proved payment.
Avery Dennison proved the product passport.
Nobody has proven the relationship.

— identity.free is building the verifiable edge.

Solution · Four passports, signed edges, one proof

Every agent action is a four-witnessed act.

Each entity gets a W3C Decentralized Identifier. Each "passport" is a profile of Verifiable Credentials 2.0. Each edge — employed-by, delegated-by, owned-by, produced-by, authorized-for — is itself a signed, scope-attenuated, revocable credential.

did:human:*

Human Passport

Proof-of-personhood, KYC/AML status, roles. Selective disclosure — prove "the agent is authorized" without revealing the principal.

did:org:*

Organization Passport

eIDAS 2.0 LPID, GS1 GLN, LEI, KYB status. The legal-person anchor that lets every downstream credential carry liability.

did:product:*

Product Passport (DPP)

GS1 Digital Link → DID → ESPR-compliant VC bundle. Manufacturing, sustainability, supply-chain, and ownership claims as verifiable credentials.

did:agent:*

AI Agent Passport

Architecture, version, training-data lineage, risk classification per the EU AI Act. Operator delegation chains. Compatible with MCP, A2A, AP2, ERC-8004.

Issue passports

W3C DID and Verifiable Credential passports for humans, organisations, products, and AI agents.

Sign the edges

Relationship credentials: employed-by, delegated-by, owned-by, produced-by, authorized-for.

Enforce scope

Expiry, limits, revocation, and policy evidence travel with every agent action.

Verify by SDK

Platforms can prove: this agent was allowed to do this action, for this entity, at this time.

Why Now · Regulatory Cliff

The trust gap moved from research to budget.

Agentic AI is entering financial workflows while Europe implements its densest regulatory wave in a generation — DORA, PSD3, FiDA, the AML package, and the EU AI Act. Institutions need infrastructure, platforms, and guardrails that make AI-native finance safe enough to deploy.

EU REGULATORY TIMELINE · 2026 → 2027

19 JUL 2026

ESPR Article 13 — Commission DPP Registry must exist; per-product unique identifiers (ISO/IEC 15459) signed by issuers.

02 AUG 2026

EU AI Act Articles 50 + Annex III — Transparency obligations; high-risk systems traceability; Article 14 human-oversight reconstructability.

31 DEC 2026

EUDI Wallet — Each Member State offers one certified wallet. LPID legal-person identity becomes operational.

30 DEC 2026

EUDR — Geolocation due-diligence statements for 7 commodities, large/medium operators.

18 FEB 2027

Battery Passport — EV/industrial >2 kWh batteries: unique-item DPPs mandatory.

END 2027

GS1 Sunrise 2027 — Retail POS globally accepts GS1 Digital Link 2D QRs.

AGENT-PROTOCOL CONVERGENCE · 2025 → 2026

APR 2025

Google A2A launched. By April 2026: 150+ supporting orgs; donated to Linux Foundation.

MAY 2025

W3C VCDM 2.0 reaches Recommendation. DID Core 1.1 Candidate Recommendation March 2026.

SEP–OCT 2025

Visa TAP, Mastercard Agent Pay, Google AP2, Stripe ACP — all launch with 100+ partners each.

DEC 2025

MCP donated to Linux Foundation. Catena Labs raises $18M (a16z) for agent payments.

2025–26

Agentic tools move from copilots to workflow execution across coding, commerce, support, and operations.

identity.free gives regulated teams an audit-ready answer to the one question they cannot hand-wave: was this agent allowed to do this action?

Wedge · Developer-first, compliance-funded

Start where agents already act. Expand where audit budgets already exist.

Months 0–9

Developer and security wedge

Sign AI-agent code commits and tool calls with human, employer, and policy authority.

Sub-week integration
Security team buyer
Forensic ledger by default

Months 9–18

Regulated finance pilots

Attach the same delegation chain to fintech operations, support agents, and agentic commerce flows.

Auditability for AI workflows
KYC/KYB and AML familiarity
Fintech design partners

Months 18+

Compliance control plane

Turn repeated authority patterns into reusable policy templates and evidence modules.

Banks and fintech platforms
Marketplaces and commerce rails
Regulator-ready evidence export

Market · Sized to the regulatory cliff

Priced by mandates, not by hype.

Identity infrastructure, non-human-identity governance, product-passport tooling, and agent-economy rails converge on the same regulated buyer over 2026–2030. All forward numbers are hypothesis-marked and re-based on pilot data.

TAM by 2030 · category sum

$113B+

Identity infrastructure ($30B+) · non-human-identity governance ($25B) · DPP / product-passport tooling ($8B) · agent-economy rails ($50B+, a16z / Bessemer thesis range).

Working thesis: 1–3% blended capture of regulated EU agent + DPP issuance — validated bottom-up through pilots.

SAM · 2026–28 priority pools

Enterprise developers shipping with coding agentsintegration-led wedge

~12M

EU-issued Digital Product Passports per yearESPR Art. 13 mandate

~95M

EU SMEs in priority ESPR sectorscompliance-funded buyers

~900K

Battery passports per year after Feb 2027unique-item DPPs, EV + industrial

~6M

Directional estimates: CIRPASS-2 baseline forecasts · active-developer telemetry extrapolation · EUR-Lex SME directives. Pinned to pilot data before any Series A claim.

Business model · Bottom-up adoption, top-down budget

Developers adopt free. Compliance pays.

Land

Developer plan

Free, usage-metered signing for agent commits and tool calls. The wedge that seeds the policy graph.

FREE + USAGE

Sub-week integration
Viral, bottom-up adoption
Highest crypto-tolerance buyers

Expand

Team SaaS

Policy orchestration, delegation management, and audit-log retention for security and platform teams.

$499–$4,999 / MO

Per-agent scope policies
Revocation and expiry control
Forensic replay on demand

Anchor

Compliance evidence

Regulator-ready evidence modules per sector for banks, fintechs, marketplaces, and agent platforms.

$25K–$250K ACV

Audit-as-API export
EU AI Act + AML mapping
Reusable policy templates

ACV BANDS ARE HYPOTHESES.

Competition · Single-Node Worlds

Every leader sells one vertex. We sell the edges.

Worldcoin has no org or product graph. Catena Labs has no products. Astrix has no consumer side. Avery Dennison has no agent stack. The signed relationship between parties is unclaimed.

Built by a fintech founder who lived the compliance pain.

Slava Solodkiy

Founder · CEO

12 years across fintech, identity, and regulated infrastructure. Founded ArivalBank, a compliance-first digital bank, and built and invested across fintech as sreda.ventures (Life.SREDA, an early fintech-only VC) — working through KYC, AML, cross-border onboarding, and auditability from the operator side. Author of the long-running Money of the Future research series. Member of the NVIDIA Developer Program and NVIDIA VC Alliance.

identity.free is the focused venture layer of that lived problem: make AI-agent authority reusable, verifiable, and safe enough for regulated finance.

Regulated scars

Hands-on familiarity with banking operations, compliance evidence, onboarding friction, and institutional risk tolerance.

Identity depth

Long-running research arc around digital identity, reusable credentials, and trust infrastructure.

AI-native timing

Clear wedge into autonomous agents as financial institutions move from pilots to production workflows.

Standards-aligned

OpenID Foundation member; built on W3C DID and Verifiable Credentials 2.0 from day one.

Company

UK entity incorporated April 2026; pre-seed stage.

Product

DID / VC schemas and delegation protocol in active build.

Credibility

Innovate UK grant application on verifiable AI for regulated finance, submitted June 2026.

Standards

OpenID Foundation member; built on W3C DID + VC 2.0.

The Ask

Help us turn agent authority into finance infrastructure.

Pre-seed · Next 90 days

Now building

Ship DID/VC schemas, delegation protocol, and SDK demo
Run 5–8 regulated-finance design-partner conversations
Secure 2 pilot design partners; validate first use case and pricing
Publish technical documentation

Design partners · banks, fintechs, marketplaces, and AI-agent platforms that need scoped authority, revocation, and audit evidence for agent workflows.

Pre-seed investors · a focused narrative around agent authority, compliance evidence, and the EU regulatory window.

Standards collaborators · OpenID, W3C DID/VC, and agent-protocol communities.

WORLDCOIN proved humanity. ENS proved naming. STRIPE proved payment. AVERY DENNISON proved the product passport.

IDENTITY.FREE proves the relationship: every AI-agent action with scoped authority, revocation, and audit evidence.

Deck