Model ListMOSS-VL-Realtime

MOSS-VL-Realtime

BetaMultimodal understanding

MOSS-VL-Realtime continuously receives JPEG, PNG, or WebP image frames over one WebSocket connection and returns incremental text answers for ongoing camera or screen understanding.

View API reference

ENDPOINT

WebSocket /v1/realtime

Current request path

Input

Continuous image frames + text

Content accepted by the request

Output

Incremental text

Content returned on success

Model overview

Realtime video understanding model, Beta. Use moss-vl-realtime-1.0 in the model connection query parameter; to pin a version, use moss-vl-realtime-1.0-2026-09-09.

Billing

Free for a limited time.

0 credits View full pricing

Modality support

TTextInput and output

AAudioNot supported

IImageInput only

VVideoNot supported

Supported endpoints

Single-speaker speech/v1/audio/speechNot supported

Multi-speaker dialogue speech/v1/audio/speech/speakersNot supported

Audio transcription/v1/audio/transcriptionsNot supported

Voice design/v1/audio/voice/generationsNot supported

Multimodal understanding/v1/responsesNot supported

Realtime video understanding/v1/realtimeSupported

Capabilities

PerformanceHigh

Inference speedRealtime interaction

Streaming Supported

Async Not supported

Bidirectional WebSocket event protocol
JPEG, PNG, and WebP image frames
Silence, interruption, and multiple answer segments
Session-level token usage reporting
No complete video files, audio input, or speech output

Versions and snapshots

Model IDmoss-vl-realtime-1.0

Snapshotmoss-vl-realtime-1.0-2026-09-09Latest