Open Standard

Context That Travels With You

Set your preferences once. Compatible AI services can adapt — in real time where integrations support live updates.

See It In Action Learn How It Works
0 Conformance Vectors
0 Protocol Layers
0 Extensions
0 SDK Languages

Set It Once. Carry It Everywhere.

One profile, compatible platforms — your preferences can travel with you

Your context can stay current — change once, participating services adapt

AI that knows your situation — and responds accordingly

How It Works

Define your context once. A compact token carries it to compatible services.

01 · You Define Defined

Your context

learn_guitar · beginner
goallearn guitar
scheduletight (shift work)
energylow today
stylehands-on learner

Your preferences, constraints, and current state.

02 · It Travels Signed

VCP token

~200 bytes · portable
encodeone token
inject()→ every service
reachmusic · learning · health

One compact token, every compatible service.

03 · Services Adapt Adapted

Live adaptation

quiet fingerpicking free tier
15-min sessions after shifts
nylon-string guitar within budget

Each service reads what it needs and adjusts.

One Token. Compatible Platforms.

Your preferences, constraints, values, and state — encoded once for compatible services.

3 Powerful Layers

Every VCP token carries three layers of context, each serving a distinct purpose.

Constitutional

Rules, boundaries, and constitutions that govern AI behavior. Hard constraints set by you that AI platforms should respect.

Rules Boundaries Constitutions

Situational

Your current role, environment, and platform context. Are you at work or home? Learning or creating? Context switches accordingly.

Role Environment Platform

Personal

Your preferences, mood, energy, and state right now. A living snapshot of who you are in this moment — updated in real time.

Preferences State Mood
Topology · I-T-S-A-M-E + Registry

Seven concepts, one protocol surface.

L1 Identity naming, namespaces, registry, encoding WHO?
L2 Transport signed bundles, ed25519, MCP delivery HOW?
L3 Semantics CSM-1 grammar, personas, traits WHAT?
L4 Adaptation context encoding, state, decay WHEN?
L5 Messaging inter-agent exchange, escalation TO WHOM?
L6 Economics fiduciary constraints, auth, governance WHO PAYS?
REG Registry canonical storage, resolution, audit WHERE?
IDENTITY TRANSPORT SEMANTICS ADAPTATION MESSAGING ECONOMICS REGISTRY
Protocol layer topology

Trust, End to End

Signed at the Source. Verified on Arrival.

Every VCP token is cryptographically signed, verified before it touches a model, and loaded with a full audit trail.

L1 · Constitution Signed

Creative Growth

personal.growth.creative@1.0.0

ed25519 · CSM-1 · SHA-256 a4f…91c

L2 · MCP Transport Verified

Value Transport

CSM-1 · persona G:3
issuercreed-space
auditorsafety-lab
inject()→ agent
L3 · Agent Loaded

Values Loaded

directivesapplied
context encoderactive
personal statedecaying
torchready

A Better Way

Without VCP
  • Re-enter preferences on many platforms
  • AI doesn't know your current state
  • No control over your data in transit
  • Static profiles that don't reflect you
  • Unsigned instructions — no proof of provenance
  • No audit trail of what the AI was told
With VCP
  • Set once, carry to compatible services
  • Real-time state awareness
  • Privacy-first, user-controlled
  • Living context that evolves with you
  • Cryptographically signed & versioned — not copy-pasted text
  • Every injection logged & replayable

Use Cases · Where It Matters

Four domains, one protocol.

Wherever values need to travel with the agent and adapt to context — personal AI, multi-agent systems, enterprise, healthcare.

Inter-agent messaging

VCP/M carries context_share, constitution_announce and escalation between agents, with severity levels and version negotiation.

See Relay demo

Personal portability

Preferences set once. Your context travels across platforms — communication style, constraints, personal state, all intact.

See Gentian demo

Private context, public behaviour

Campion's personal constraints (single parent, health appointments) shape advice without ever reaching HR or the platform.

See Campion demo

Reflective guidance

Marta weighs a family loan request. The Mediator persona uses her private context to distinguish genuine duty from guilt.

See Marta demo

See It In Action

Illustrative scenarios, contextual signals, and protocol-mediated adaptation.

01

Gentian's Guitar Journey

Learning guitar across 4 platforms with one persistent profile.

Portability
Try Demo
02

Campion's Corporate Training

Dual profiles that switch automatically with your role.

Adaptation
Try Demo
03

Marta's Responsibility Journey

Values and real-time state shaping AI guidance moment to moment.

Liveness
Try Demo
04

Relay: Agent Handoff

Values context traveling between agents in a procurement chain.

Coherence
Try Demo
Explore All Demos
token.vcp
v3|usr:gentian|key:abc123
con:guitar-learning
per:visual-learner
goal:fingerpicking
flags:dyslexia-friendly
priv:medical=none
Compact Encoding

Tiny Tokens, Rich Context

VCP tokens pack your entire context into a compact, portable format. Preferences, constraints, current state — all traveling in a few hundred bytes.

  • Human-readable format
  • Privacy-preserving layers
  • Real-time updates
Learning
Work
Health
Shopping
You
Cross-Platform

One You, Compatible Services

Your context can follow you across participating AI services. No more re-entering preferences, re-explaining constraints, or starting from scratch.

  • Zero re-entry
  • Automatic adaptation
  • Platform-agnostic

Build Your Own Context

Create, inspect, and experiment with VCP tokens in real-time. Toggle constraints, change preferences, and watch the compact token update instantly.

Open Playground

Built for Developers

One Spec. Three SDKs.

Drop the governance layer into your stack in Python, TypeScript, or Rust. Authorize tool calls with cryptographic proof, negotiate capabilities, and carry context — all tracking VCP v3.1.

  • Ed25519 signing, SHA-256 content hashes
  • VCP-Hello / VCP-Ack capability handshake
  • Personal State with exponential & linear decay
  • Schulze consensus & Torch session handoff
Read the Docs View on GitHub
from creed_sdk import create_client

client = create_client(api_key="crd_live_...")

# Authorize a tool call through the governance layer
result = await client.decide(
    tool_name="send_email",
    arguments={"to": "user@example.com", "subject": "Hello"},
    on_allow=lambda d: print(d.decision_token),
    on_deny=lambda d: print(d.reasons),
)

Want to Understand the Details?

How VCP Works Free Course (~4 hours) Technical Documentation