Private Face API Product

NEVAFACE

Face Detection, Embedding and Search API

Simple.

Private.

Customer-owned data.

Built as a sellable API product, not a messy demo stack. Customers keep their own embeddings, and we do not store face images or embedding data.

Why NEVAFACE?

✓ No database lock-in
✓ No face images stored
✓ Customer owns embeddings
✓ Simple 3 endpoint architecture
✓ Ready for MongoDB Vector Search
✓ Private deployment available
Fast detect -> embed -> search sequence
Private API product ready for commercial rollout
Lawful, privacy-compliant implementation focus
No vendor lock-in around vector storage

Processing Flow

Private API

1. Detect

Validate the image and count faces.

↓

2. Embed

Turn the selected face into a vector.

↓

3. Search

Compare against customer-provided embeddings.

Three Endpoints

One simple commercial structure

Detect first, embed second, search last. A straightforward flow that is easier to sell, integrate, and deploy privately.

Step 1

Face Detection

/face/detect

Checks whether the image contains faces before you spend compute on embedding or matching.

  • • Returns face count and bounding boxes
  • • Useful for camera checks and human presence validation
  • • Best entry point before all downstream face workflows
{
  "success": true,
  "faceCount": 2,
  "faces": [
    {
      "faceId": 1,
      "confidence": 0.998
    }
  ]
}

Step 2

Face Embedding

/face/embed

Turns a detected face into a numeric vector that your customer can store in their own database.

  • • 512-dimension embedding output
  • • Works with PostgreSQL, Redis, MongoDB and custom pipelines
  • • No face images or embeddings are stored on our side
{
  "success": true,
  "model": "insightface",
  "embeddingDimensions": 512,
  "faceDetected": true
}

Step 3

Face Search

/face/search

Compares a fresh face against customer-provided embeddings and returns the closest match.

  • • Threshold-based matching
  • • Ideal for private access systems and internal identity checks
  • • Customer controls the candidate embedding list
{
  "success": true,
  "matched": true,
  "checkedCount": 24,
  "bestMatch": {
    "id": "user_1",
    "similarity": 0.912
  }
}

Use Cases

Built for teams that already own their data

Private member verification
Kiosk and camera onboarding checks
Employee or visitor lookup systems
Internal face-based matching products

Sales Positioning

Website promo first, paid API second

NEVAFACE can live as a clean product page on your website while access is provisioned privately. That lets you market the product now and scale API distribution later.

Suggested rollout:
1. Promote NEVAFACE on-site.
2. Collect interest and validate demand.
3. Open private access for selected customers.
4. Publish on external API marketplaces later.

Privacy Notice

Built for privacy-sensitive deployments

NEVAFACE is intended for lawful and privacy-compliant use cases only. Users are responsible for obtaining all required permissions and complying with local privacy regulations.

We do not retain face images, embeddings, or customer biometric datasets. Customers keep control of their own stored vectors and search candidates.