API Endpoints¶
Overview of all available Moveris API (v2) endpoints.
In plain terms
Moveris offers real-time and async paths. For frame-based integrations, use fast-check, fast-check-stream, or fast-check-crops. For pre-recorded full videos, use video/detect (submit once, then poll). Frame count depends on the model (for example, 10 for mixed-10-v2, 30 for mixed-30-v2).
Base URL¶
Health Check¶
| Method | Endpoint | Description |
|---|---|---|
GET | /health | Check service status and loaded models |
View Health Check documentation
Fast Check¶
| Method | Endpoint | Description |
|---|---|---|
POST | /api/v1/fast-check | Server-side face detection (frame count matches model) |
POST | /api/v1/fast-check-stream | Send frames one-by-one; verdict when all frames received |
POST | /api/v1/fast-check-crops | Pre-cropped face verification (faster) |
POST | /api/v1/live-check | V3 batch path with client-supplied MediaPipe landmarks per frame |
V2 Upload-First Pipeline¶
| Method | Endpoint | Description |
|---|---|---|
POST | /api/v2/upload | Buffer frames (or video) during capture — no inference |
POST | /api/v2/fast-check | Run inference by session_id after upload completes |
POST | /api/v2/live-check | V3 live-check variant — upload frames with landmarks first |
View V2 pipeline documentation
v1 vs v2
v1 endpoints send all frames in the inference request. v2 uploads frames first, then calls a lightweight check with only session_id. Both coexist; choose based on your capture flow. The React SDK useLivenessV2 hook targets the v2 pipeline.
Video Detect (Async)¶
| Method | Endpoint | Description |
|---|---|---|
POST | /api/v1/{tenant_slug}/video/detect | Submit full video (file or URL) for async processing |
GET | /api/v1/{tenant_slug}/video/detect/{submission_id} | Poll status and fetch final verification result |
When to use video/detect
Use this flow when you already have a full recorded video and want async processing. You submit once and poll by submission_id until completion.
View Video Detect documentation
Endpoint Comparison¶
| Endpoint | Use Case | Face Detection | Latency |
|---|---|---|---|
/fast-check | Standard integration | Server-side | ~245ms |
/fast-check-stream | Real-time streaming | Server-side | ~245ms |
/fast-check-crops | Maximum performance | Client-side | Fastest |
/live-check | V3 with client landmarks | Client + server | ~245ms |
/api/v2/upload + /api/v2/fast-check | Upload-first flow | Server-side | Varies by upload chunks |
Choose the right endpoint
- Use fast-check for simple integrations where you send all frames at once (count depends on model)
- Use fast-check-stream when capturing frames in real-time from a camera
- Use fast-check-crops for the fastest processing when you can do face detection client-side
Recommended models
Use mixed-30-v2 (Balanced) for most integrations. See Models overview for frame counts and performance characteristics.