SightRadar
Face recognition API, Rekognition-compatible, pay per photo
In their words
SightRadar is a face recognition API for teams that need face search, verification and detection at scale, without AWS Rekognition's pricing. It is Rekognition-compatible: point the AWS SDK (boto3 and others) at SightRadar's endpoint and your existing code keeps working.
What you get:
- Face search: find one person across thousands of photos from a single selfie (event photos, media libraries, photo delivery)
- 1:1 verification: compare two faces with a similarity threshold you choose
- Face detection with confidence scores you can threshold on
- Rekognition-compatible API: CreateCollection, IndexFaces, SearchFacesByImage, CompareFaces, DetectFaces and more
- Photos with no faces return valid empty results, not errors
- Per-photo billing, never per face: batch indexing from $0.00062 per photo, real-time $0.00093 per photo
- Privacy by design: GDPR-ready with a DPA, face vectors encrypted per account, never used for model training
- SDKs on npm and PyPI, plus a Rekognition shim for Python
Free to start with trial credits, no credit card required. Pay-as-you-go prepaid credits, no subscription and no minimums.
Fact file
Developer tool- API
- AWS Rekognition
- boto3
- Batch indexing
- $0.00062 per photo
- Real-time index
- $0.00093 per photo
- Storage
- $6 per million stored faces per month
What is SightRadar?
High-accuracy face recognition and verification run through an API built for engineering teams seeking an alternative to AWS Rekognition. SightRadar operates as a drop-in replacement, letting developers keep existing SDK implementations by pointing them to a single new endpoint without rewriting underlying integration logic.
Rekognition-compatible endpoint migration
The service supports standard request and response shapes, meaning existing code written for AWS Rekognition continues working with minimal configuration. By swapping the endpoint URL in the boto3 client or other supported SDKs, applications execute core operations like CreateCollection, IndexFaces, SearchFacesByImage, SearchFaces, CompareFaces, DetectFaces, ListFaces, DeleteFaces, ListCollections, and DeleteCollection without changing how data flows into the system.
Core face operations and billing structure
Four main operations handle face processing tasks. Indexing detects and stores face embeddings in collections, searching finds matching photos from a single selfie, 1:1 comparison verifies identity similarity, and detection locates faces with confidence scores without persistence. Billing is assessed per photo processed rather than per face detected. Photos containing zero faces return valid, successful results instead of errors. Pricing tiers include real-time indexing at $0.00093 per photo, batch indexing at $0.00062 per photo, and storage priced at $6 per million stored faces per month.
Data handling and privacy controls
Tenant isolation ensures that face vectors from one account are never cross-searchable with another. Detect and compare operations process images exclusively in memory without persisting them on disk. Data is encrypted in transit, and access is managed via scoped API keys formatted with specific prefixes. The platform prohibits using customer data for training general-purpose models, maintaining a GDPR-ready posture for lawful and consented applications.