NeuraRelay

Neura Relay governs agent actions before execution

Authority before an agent changes anything

Diagram showing an agent action passing through Neura Relay Authority checks and returning a Decision Receipt
See how Relay worksTry Playground

Where Relay fits

Between agent intent and downstream execution

01

Agent proposes action

A runtime prepares a customer, data, content, payment, or tool action

02

Relay checks context

Identity, authority, evidence, policy, risk, and intent become explicit

03

Receipt comes back

The decision explains proceed, revise, review, or stop before execution

04

Runtime stays yours

Your product decides how execution continues after reading the receipt

Open Authority path

Resolve a Proposed Action

Example 1 of 4

Rotating examples

Use arrows for all four examples. Playground runs full review.

Decision Receipt

Decision
Ready

Ready to proceed

Confidence
High

Recommended next step

Developer may continue to execution

Reason

The action has approved content, a bounded recipient, and no irreversible downstream system change

High confidence to proceed, score range 78–92

Add Relay at the action boundary

One resolve call returns a Decision Receipt before your runtime acts

Integration path
  1. 01

    Build Action Card

    Attach actor identity, proposed action, affected object, evidence refs, and requested outcome

  2. 02

    Check factors

    Relay resolves the proposal into visible decision factors before downstream execution

  3. 03

    Use the receipt

    Choose proceed or revise stop or route for human review before execution

Resolve call
Example
TypeScript
const response = await fetch("/api/resolve", {  method: "POST",  headers: { "Content-Type": "application/json" },  body: JSON.stringify({ action_card: actionCard })});
const response = await fetch("/api/resolve", { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify({ action_card: actionCard }) }); const { decision_receipt: receipt } = await response.json(); if (receipt.decision === "proceed") { execute(actionCard); } else { queueHumanReview(receipt); }
Open GitHub exampleOpen Protocol reference
Core example

Direct Action Card to Decision Receipt

MCP proof

Optional protected adapter path

Blocked action

High-risk action routes to stop

Before agents act

Action Card in. Decision Receipt out.

A 63-second film introducing the governed path from Action Card to Decision Receipt.

Open on YouTube

The stack behind governed action

Relay checks decisions, Registry identifies agents, Protocol defines the message structure

Developer operating stack
Decision Infrastructure

Neura Relay

Check identity, authority, evidence, policy, risk, and flow before execution

Open Authority path
Builder Registry

Neura Registry

Register agents, declare capabilities, issue versions, and keep identities ready for Relay reference

Open Builder
Reference Model

Neura Protocol

Define Action Card, decision factor, receipt, trace, lifecycle, validation, and conformance structure

Open Reference

Questions before production

Identity, receipts, execution boundaries, and optional MCP compatibility

Integration FAQ

Relay gives AI products an authority layer before an agent changes a customer record, sends a reply, triggers a workflow, or calls a tool. Your product keeps execution; Relay returns the governed receipt.