abbe

Media orchestration, one conversation away

abbe — Your ai based broadcast engineer

Design, orchestrate, and monitor broadcast and AI media streams in real time. Powered by natural language MCP and a visual Web UI.

Visual Web UI

Build with your eyes and hands.

Visual pipeline builder

Arrange and wire media stages on a precise hex canvas.

Live stream telemetry

See bandwidth, RTT, packet loss, and bitrate as streams run.

One-click provisioning

Price, deploy, and follow every lifecycle step from one workspace.

Model Context Protocol

Build from the tools you already use.

  • Build and reconfigure pipelines in natural language
  • Price and deploy through the same audited actions as the Web UI
  • Connect from Cursor, Claude Desktop, OpenCode, or Codex
Claude Desktopvia abbe MCP
User prompt: abbe, create an srt fanout of 20 locations from our primary ingress feed
Claude response through abbe MCP: I’ve resolved the primary feed into 1 SRT ingress and 20 SRT listener egress endpoints distributed across the requested edge regions. Target latency is 120 ms per destination, with an estimated 256 Mbps aggregate throughput and 625 credits per hour. The draft is ready for endpoint review before pricing and deployment.
CursorClaude DesktopOpenCodeCodex

Browser UI in action

From design to live signal.

abbe visual pipeline Builder with an SRT fanout canvas and Stage Catalog
Visual BuilderReady to configure
abbe browser workspace showing live media components and telemetry-ready cards

Live operations workspace

Inspect active media capabilities and move directly into monitoring and control.

Expanded SRT Fanout Distribution pipeline card with diagram and details

Pipeline details

Open a pipeline card to inspect its diagram, identity, description, and next action.

Stage Inspector editing an SRT egress stage

Stage Inspector

Configure customer-owned stage settings while provisioned values stay clearly locked.

Ready-to-run examples

Start with a proven pipeline.

Choose a production-shaped template, inspect its real vector netlist, then let abbe guide the launch.

SRT Fanout Distribution

Terminate one SRT input, replicate the flow, and deliver independent SRT outputs without duplicating the source contribution.

5 stagesSRTFanout
SRT IngestSRT IngestSRT FanoutSRT FanoutSRT EgressSRT EgressSRT EgressSRT EgressSRT EgressSRT Egress

Live Speech Translation & Captioning

Transcribe live program audio with an AI speech stage, package the result as timed events, and deliver captions to downstream systems.

4 stagesSRTSpeech AI
SRT IngestSRT IngestSpeech to Text (Azure Speech)Speech to Tex…Event EmitterEvent EmitterEvent SinkEvent Sink

AI Vision Stream Inspection

Inspect inbound video with an AI vision model, compose its detections onto the program, and deliver the annotated stream over SRT.

4 stagesSRTVision AI
SRT IngestSRT IngestObject Detection (AWS Rekognition)Object Detect…Video OverlayVideo OverlaySRT EgressSRT Egress

Low-Latency Broadcast Edge

Receive an SRT contribution, stabilize timing at the edge, and deliver a low-latency output with real-time operational visibility.

3 stagesSRTLow latency
SRT IngestSRT IngestTiming BufferTiming BufferSRT EgressSRT Egress

Your next media pipeline starts here.

Open the visual workspace or describe the outcome to abbe through any MCP-compatible tool.