Loading DeepShield AI...
Loading DeepShield AI...
Integrate sub-second deepfake verification and multi-modal neural forensic analysis into your media workflows and verification desks.
Instant sub-50ms cryptographic cache lookup by video SHA-256 fingerprint.
Generate direct Azure Blob Storage SAS tokens for zero-proxy ephemeral upload.
Real-time Server-Sent Events (SSE) telemetry tracking the 5 worker stages.
Checks the edge Redis cache for previous forensic analyses matching the video's 64-character hex SHA-256 fingerprint.
curl -X GET "https://deepshield.video/api/lookup?hash=e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" \ -H "Accept: application/json"
{
"cached": true,
"hash": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
"result": {
"jobId": "job_01h7x9k2m",
"verdict": "authentic",
"fakeProbability": 0.08,
"confidenceScore": 0.94,
"methodAttribution": {
"faceSwap": 0.02,
"lipSync": 0.04,
"diffusion": 0.01,
"audioClone": 0.01
},
"analyzedAt": "2026-10-10T02:15:30.000Z"
}
}Integrate the pre-upload client hash query to avoid transferring media that has already been verified.
import { sha256 } from "hash-wasm";
async function verifyVideoBeforeUpload(videoBlob: Blob) {
// 1. Compute client-side SHA-256 hash in browser memory
const buffer = await videoBlob.arrayBuffer();
const hash = await sha256(new Uint8Array(buffer));
// 2. Query DeepShield instant edge cache
const res = await fetch(`https://deepshield.video/api/lookup?hash=${hash}`);
const data = await res.json();
if (data.cached) {
console.log(`Instant verdict: ${data.result.verdict} (${data.result.fakeProbability * 100}% synthetic)`);
return data.result;
}
console.log("Cache miss. Video must be uploaded via /api/upload-url flow.");
}