reticle

Hardware attestation SDK for JavaScript and TypeScript. Cryptographically verifies AMD SEV-SNP, Intel TDX, and NVIDIA GPU attestations from an attestation server — runs in Node.js, Bun, Deno and the browser via WASM.

Install

npm install @premai/reticle

Quick start

import { ClientBuilder } from "@premai/reticle";

const client = await new ClientBuilder("https://attestation.example.com")
  .build();

// End-to-end attestation: discovers modules, generates nonces,
// verifies cryptographic signatures and certificate chains
const result = await client.attest();

console.log("CPU:", result.modules().cpu());   // CpuModule.Sev or CpuModule.Tdx
console.log("GPU:", result.modules().gpu());   // GpuModule.Nvidia or undefined

API

ClientBuilder

Creates and configures a Client instance. build() is async.

const client = await new ClientBuilder(url)
  .with_authorization("Bearer <token>")  // optional: set Authorization header
  .with_kds(new Kds(kdsUrl))             // optional: custom AMD KDS cache
  .with_pcs(new Pcs(pcsUrl))             // optional: custom Intel PCS cache
  .with_policies_url(policiesUrl)        // optional: custom OPA policies index URL
  .build();

Client

High-level (recommended)

These methods handle nonce generation, request, and full cryptographic verification in a single call:

| Method | Returns | Description | |---|---|---| | client.attest() | Promise<AttestResult> | Full attestation of all available modules (CPU + GPU) | | client.attest_sev() | Promise<void> | AMD SEV-SNP attestation only | | client.attest_tdx() | Promise<void> | Intel TDX attestation only | | client.attest_nvidia() | Promise<ResponseHeaders> | NVIDIA GPU attestation only |

To pass custom query parameters, call set_query before attesting:

const query = new QueryParams()
  .with("model", "my-model")
  .with("version", "1.0");

client.set_query(query);
await client.attest();

Low-level

These methods fetch and parse attestation evidence (discouraged) — useful when you need to inspect raw data or implement custom validation:

| Method | Returns | Description | |---|---|---| | client.request_modules() | Promise<Modules> | List available attestation modules on the server | | client.request_sev(nonce: SevNonce) | Promise<ParsedAttestation> | Fetch raw SEV-SNP attestation report | | client.request_tdx(nonce: TdxNonce) | Promise<TdxQuote> | Fetch raw TDX quote | | client.request_nvidia(nonce: NvidiaNonce) | Promise<NvidiaAttestResult> | Fetch raw NVIDIA EAT attestation result |

AttestResult

Returned by client.attest().

| Method | Returns | Description | |---|---|---| | result.modules() | Modules | CPU and GPU module info | | result.headers() | AttestHeaders | HTTP response headers from attestation requests |

Modules exposes:

• .cpu() → CpuModule (CpuModule.Sev or CpuModule.Tdx)
• .gpu() → GpuModule | undefined (GpuModule.Nvidia or undefined)
• .has_gpu() → boolean

AttestHeaders exposes:

• .cpu() → ResponseHeaders | undefined
• .gpu() → ResponseHeaders | undefined

QueryParams

Pass custom query parameters to the attestation server. The nonce key is reserved and will throw if used.

const query = new QueryParams()
  .with("model", "my-model")
  .with("version", "1.0");

client.set_query(query);

Low-level sub-module usage

Lower-level types are available as top-level named exports:

import {
  fetch_keychain,
  EATToken,
  Kds,
  NvidiaNonce,
  SevNonce,
} from "@premai/reticle";

// NVIDIA: manual token parsing and verification
const keychain = await fetch_keychain();
const nonce = NvidiaNonce.generate();
const token = EATToken.parse(rawJwt);
const claims = token.verify(keychain, nonce);

// AMD SEV-SNP: manual certificate chain verification
const kds = new Kds("https://kds.example.com");
const sevNonce = SevNonce.generate();
const chain = await kds.fetch_certificates(attestation);
attestation.verify(chain, sevNonce);

Memory management

WASM objects are not garbage-collected automatically. Call .free() when done, or use using (TypeScript 5.2+) for automatic disposal:

using client = await new ClientBuilder(url).build();
using result = await client.attest();
// automatically freed at end of scope

Examples

• Bun — minimal CLI attestation
• Vite — browser UI with attestation status

License

See LICENSE.