Identity is fragmented
Digital identity is scattered across platform databases, passwords, government systems, enterprise directories, and proprietary login providers.
A clean-slate trust layer for the agentic age
The internet is becoming intelligent.
Its trust layer is not.
ActiveChain is post-quantum from day one: a protocol for verifiable identity, programmable authorization, smart contracts, and accountable AI.
One global state where specialized actors can compute, build, prove, store, and serve—without becoming authorities users must trust.
Protocol design · Pre-testnet · Open research and implementation
AUTHORITY DOWN · EVIDENCE BACK
01Human · organization · AI agent
02Principal
03Credential + capability + policy
04Object or compute job
05Execution
06Proof or receipt
07Finalized global state
08Independent verifier
01Authority is explicit
Every action must be authorized by a scoped, machine-verifiable policy.
02Computation carries evidence
Important results arrive with a proof or an explicitly labeled assurance receipt.
03Specialization without trust
Infrastructure providers can specialize without gaining control over validity.
01 / Why now
The foundations were built for another world.
The internet moved information between machines. Blockchains transferred scarce digital assets between keys. Neither was designed for autonomous agents that negotiate, spend, create, and make decisions on behalf of people.
Authorization is implicit
Possession of a key, API token, or wallet often grants excessive authority. Logs explain a breach only after it has happened.
AI agents are over-permissioned
Most agents operate with reusable credentials, opaque tool access, poorly defined budgets, and no protocol-level distinction between a suggestion and an authorized action.
Privacy and verification are afterthoughts
Systems expose data by default and add privacy, provenance, and verifiable computation later through layers with different trust assumptions.
We are trying to run an agentic, machine-mediated civilization on foundations that never modeled identity, authority, privacy, or accountable computation.
02 / First principles
Some foundations cannot be repaired with another layer.
ActiveChain starts from scratch because previous systems taught us precisely where the deepest constraints are.
Explore the researchRETROFITTED STACK
Base ledger
Account abstraction
Identity provider
Bridge / rollup
Privacy / oracle
AI service
Governance multisig
Each layer introduces a new interface, operator, failure mode, and trust assumption.NATIVE FOUNDATION
PrincipalCredentialCapabilityPolicyObjectJobProof / Receipt
One semantic foundation for identity, authority, state, computation, privacy, and evidence.A signature proves that someone controlled a key. It does not prove who they were, what they were allowed to do, or why the action should be accepted.
03 / Native primitives
A protocol that understands authority.
Every state transition is expressed through seven native concepts—available to wallets, contracts, organizations, autonomous agents, compute workers, and privacy proofs.
01
Principal
A persistent identity for a person, organization, device, service, pseudonym, or AI agent.
VALID TRANSITIONAvailable data + Authenticated principal + Authorized action + Correct execution + Satisfied evidence policy
04 / What this makes possible
A safer foundation for digital agency.
Autonomous agents with enforceable limits
An agent receives a time-limited capability defining exactly which tools, data, budgets, counterparties, and actions it may use.
Book travel up to €500/day · never initiate a bank transfer
Private, reusable digital identity
Prove age, membership, professional status, authorization, uniqueness, or non-revocation while revealing only what the transaction requires.
Prove the policy. Keep the record.
Proof-carrying smart contracts
Explicit, resource-safe objects execute in parallel while effects remain atomically composable within one global state.
Validity is a property of the transition
Verifiable AI and computation
Bind an output to an exact model, runtime, input commitment, random seed, data authorization, and execution method.
Model · input · seed · evidence
Post-quantum security from the foundation
The core security path requires no classical elliptic-curve primitive to authorize state, finalize consensus, or verify execution.
Versioned cryptography, replaceable by design
Specialization without authority
Builders, provers, AI workers, and storage providers can perform useful work without fabricating an authorized state transition.
Many providers · one validity standard
05 / Architecture
One state. Many specialized providers. No implicit authorities.
Specialize the work. Decentralize the trust.
01Principals & policyHumans · organizations · devices · services · agents
02Data availabilityMulti-producer publication · erasure coding · snapshots
03Consensus & orderingPost-quantum BFT · protected transactions · ordering
04Objects & smart contractsResource-safe objects · parallel execution · atomicity
05AI & compute jobsModel manifests · dataset rights · evidence policies
06Proofs & finalityAuthorization proofs · recursive aggregation · final state
ORDINARY VERIFIERVerify availability, consensus, authorization, and one succinct transition proof.
No need to trust the builder, prover, AI worker, archive, or application operator.
06 / Privacy
Privacy without invisible authority.
Privacy is not a single switch. Each guarantee should be understood honestly.
01Protected ordering
Transaction details can remain encrypted until the transaction set is locked.
02Private credentials
Prove an issuer made an acceptable claim without disclosing a global identity.
03Private policies
Prove a confidential authorization policy returned permit.
04Shielded objects
Ownership, value, state, and application data can remain encrypted.
05Selective auditability
Reveal the minimum evidence to the people and institutions entitled to audit.
07 / Roadmap
From semantic kernel to first testnet.
NOW · COMPLETESemantic devnet
Canonical primitives, authority kernel, ObjectVM, credentials, policy, state tree, objects, and deterministic single-node execution.
NOW · COMPLETEComposable packages
Immutable ObjectVM packages, bytecode commitments, safe entry points, imports, and governed upgrades.
NEXT · TESTNET GATEProof-carrying execution
Execution evidence, multi-node verification, and the first adversarial testnet release criteria.
08 / FAQ
Questions worth answering early.
Is ActiveChain live today?
No. It is protocol design and open implementation, currently progressing from a deterministic semantic devnet toward its first testnet.
Is this another cryptocurrency?
No. The focus is a verifiable object ledger where identity, authorization, privacy, computation, and evidence are protocol primitives.
How can I contribute?
Read the specifications, run the local test suite, challenge the assumptions, and contribute through the public GitHub repository.
Protocol design · Pre-testnet
Build the trust layer with us.
ActiveChain is open research and implementation. Read the architecture, challenge the assumptions, and help shape a protocol for a machine-mediated world.
NEXT / ROADMAP
Architecture → reference implementation → devnet → pre-testnet
Follow progress on GitHub