XBPP.ORG
    Back to Homepage

    Definition: xBPP (Execution Boundary Permission Protocol) is the open standard for governing autonomous AI agent payments. It provides a chain-agnostic governance layer that evaluates agent spending intent against declarative policies before transactions are executed.

    Architecture of Trust

    xBPP is a stateless, payload-driven standard. It doesn't live on one chain, and it doesn't require a specific wallet. It works by bringing governance directly to the point of intent.

    HOW IT ALL CONNECTS

    TRUST IS THE PRODUCT.

    A standardized middleware layer between Agent and World

    YOUR APPLICATION

    Agent wants to pay $50 to 0xABC...

    XBPP INTERPRETER

    Action + Policy → Verdict

    RESULT

    ALLOW | BLOCK | ESCALATE

    Step 1

    Intent

    Agent generates an action request

    "Pay $50 USDC to 0xABC on Base"

    Step 2

    Interpret

    xBPP checks against active Policy

    Constraints evaluated in real-time

    Verdict

    ALLOW

    Proceed with the payment

    BLOCK

    Stop; this violates policy

    ESCALATE

    Ask a human to approve

    Key Concepts

    ConceptWhat It IsExample
    ActionA proposed payment to evaluate"Pay $50 USDC to 0xABC"
    PolicyRules defining what's allowed"Max $100/tx, $1000/day"
    VerdictThe evaluation resultALLOW, BLOCK, or ESCALATE
    StateRunning totals and history"$340 spent today"
    PostureDefault risk toleranceAGGRESSIVE, BALANCED, CAUTIOUS

    Graduated Autonomy

    Trust builds over time - automatically

    Cautious Posture

    $10 limit, full human oversight

    Aggressive Posture

    $10k limit, autonomous execution

    Graduate between postures as trust builds - without rewriting a single line of code.

    XBPP vs TRADITIONAL RAIL

    THE PARADIGM SHIFT

    HOW XBPP CHANGES EVERYTHING

    Current State
    With xBPP
    Control

    Hard-coded if (amount > 100) inside the bot

    External Policy file: max_single: 100

    Update

    Redeploy the entire agent to change a limit

    Update the Policy; agent obeys instantly

    Audit

    "Check the logs and hope"

    Cryptographic Verdict attached to every transaction

    Liability

    Blame the developer

    Audit the verdict - the Protocol enforced the Policy

    1. The Intent

    An agent prepares an x402 transaction. This could be a purchase, a swap, or a subscription payment.

    2. The Evaluation

    The transaction payload is passed through the xBPP engine. Policies are checked against current state and transaction parameters.

    3. The Verdict

    xBPP returns a signed verdict: ALLOW, BLOCK, or ESCALATE. Only 'ALLOW' produces the required signature for settlement.

    Ready to see it in action?

    Run a Policy CheckRead the Spec