Live stream
One ingest
A single live feed comes in. No second ingest, no extra cables.
- MPEG-TS
- SRT
- HLS
- RTMP
Coming soon
- SMPTE ST-2110
- JPEG-XS
Sense platform
Catch it while the show is still live
Dead air, garbled audio, caption errors, language drift, and picture faults — on every tap, without touching the program.
Universal Media TapsCapabilitiesDeploymentsIntegrations
Universal Media Taps
One ingress layer for every source
A tap is a clean, AI-ready copy of your live stream, peeled off ingest so Senses can analyze without touching the viewer pipeline. One ingest, two jobs: watchable media for humans, analyzable media for AI.
- Streaming: MPEG-TS, SRT, HLS, RTMP
- File ingest: MP4, MXF OP-1a, MOV, broadcast-grade containers
- API push: REST ingest endpoint for transcode output
- Coming soon: SMPTE ST-2110, JPEG-XS
Main pipe
Playback keeps its own pipe
Operators keep watching. Preview and on-air stay independent. AI never steals the main feed.
Taps
AI-ready copies, on by need
Turning on a Sense is what creates its tap. Audio-only AI means an audio tap and nothing more, so unused plumbing costs nothing.
- AudioClassify · Transcribe
- VideoFrames for vision models
- MetadataCaptions · SCTE-35
Senses emit findings as events on the timeline, never the raw media.
What Sense sees across audio, video, and metadata once taps are live.
Audio
Video
Metadata
Capabilities
What Sense detects, 24/7
Six detection domains run in parallel across every tap, every feed, every minute of the day. Pick a job to see it in the control room.
Catch it while the show is still live
Detection domains
Loudness & CALM compliance
Continuous ITU-R BS.1770, CALM Act, and EBU R128 monitoring per segment, break, and program — with timestamped evidence clips.
Closed caption QC
FCC Part 79 presence, timing, and accuracy checks across all tracks — for air and catch-up streams.
Audio content classification
Speech, music, silence, and ambient detection per frame. Feeds loudness, rights, and language workflows.
SCTE-35 cue detection
Splice insert, splice null, and blackout trigger detection with insertion-rate anomaly alerting.
Video QC
Black frame, freeze frame, and signal-dropout detection with frame-accurate timestamps and operator clips.
Real-time alerting
Configurable thresholds push to the operator surface, PagerDuty, Slack, and webhook endpoints — sub-second from detection.
Deployments
On-prem unmetered, cloud elastic
On-premises
Run inside your NOC. Fixed annual license with no per-stream overages, no cloud egress, and no data-sovereignty concerns.
- 24/7 unmetered processing
- Air-gapped or private-network deployment
- Hardware appliance or VM on your own infrastructure
- Predictable annual pricing
Cloud
Spin up capacity for events, overflows, and cloud-native pipelines — burst for the tentpoles, scale down between them.
- AWS, Azure, and GCP regions
- Auto-scale to thousands of concurrent streams
- Burst capacity for live events and tentpoles
- Flat per-track rate — PAYG for one-off events
Integrations
Fits your existing stack
Pre-built connectors cover the broadcast infrastructure you already run. No rip-and-replace required.
AWS Elemental
MediaLive and MediaConvert output taps
Harmonic
ViBE and Spectrum encoder output
Grass Valley
AMPP and Morpheus automation hooks
Telestream
Vantage transcoding pipeline
Imagine Communications
Versio and Selenio playout output
Webhook / REST
Any pipeline with an HTTP output endpoint
Ready to instrument your pipeline?
Create an account, size your tracks, and subscribe. We provision on-prem licenses after checkout — or book a technical walkthrough with engineering.