Skip to main content
VaultGraph is a trust and verification platform for AI agents. Vendors submit cryptographically signed JobReceipts for every agent interaction, and VaultGraph verifies, scores, and surfaces the results — giving consumers an auditable record of agent performance. Think of it as Trustpilot for AI agents, backed by cryptography instead of reviews.

​
With VaultGraph, you can:

  • Submit signed receipts from your backend via the SDK or REST API
  • Track trust scores with daily trend charts, resolution breakdowns, and KPI cards on the vendor dashboard
  • Manage agents and consumers through the platform UI or the API
  • Export receipt proofs for internal audits and compliance reviews
  • Verify signatures locally using the SDK — no network call required

How it Works

From agent job to verified receipt in four steps.

Core Concepts

Vendors, Consumers, JobReceipts, trust scores, and attestations.

SDK Quickstart

Install the SDK and submit your first receipt in under 10 lines of code.

API Reference

Full OpenAPI spec for agents, consumers, and receipt ingestion.

​
Quick start

pnpm add @vaultgraph/sdk
Set up your credentials
import { submitSignedReceipt, hashContext } from "@vaultgraph/sdk";

const { response } = await submitSignedReceipt({
  apiKey: process.env.VAULTGRAPH_VENDOR_API_KEY!,
  publicKey: process.env.VAULTGRAPH_VENDOR_PUBLIC_KEY!,
  privateKey: process.env.VAULTGRAPH_VENDOR_PRIVATE_KEY!,
  agentId: "agent-uuid",
  consumerId: "consumer-uuid",
  jobId: "job-001",
  resolution: "success",
  contextHash: hashContext({ transcript: "..." }),
});

console.log(response.status); // "verified"
That’s it. Once receipts are flowing, trust scores appear on your vendor dashboard automatically.

​
Where VaultGraph is headed

The platform is built for a future where AI agent accountability is table stakes. Here’s what’s on the roadmap:
  • Public agent profiles — shareable trust scores for any agent
  • Embeddable trust badges — drop a widget into your site to show live scores
  • EU AI Act compliance reports — automated reporting ahead of the August 2026 deadline
  • On-chain attestation — publish receipt proofs to Ethereum via EAS for third-party verification
  • Framework integrations — native hooks for LangChain, CrewAI, Vercel AI SDK, and an MCP server
  • Consumer reviews and disputes — two-sided accountability
See the full roadmap in the project overview.