Use case

AI-generated video

AI video models (and pipelines around them) emit an MP4 at a URL. rehelios imports that URL server-side, runs the full adaptive-HLS transcode, and returns an embeddable, edge-delivered URL — no client re-upload. The whole flow is one MCP tool call or one CLI command, so an agent can generate a clip and put it live autonomously.

Start free Read the docs

When to use rehelios for this

Choose rehelios when an agent or pipeline produces video files (from a generation model, a renderer, or a screen recorder) and you need them transcoded to adaptive streaming and hosted, without building upload plumbing.

  • URL import — no client-side re-upload of model output
  • Single MCP tool call (rehelios_import) for agents
  • Adaptive HLS so generated clips play on every device

How it works

  1. 1

    Generate

    Your model or pipeline produces a video file reachable at a public URL.

  2. 2

    Import by URL

    Call rehelios_import (MCP) or POST /v1/videos/import — rehelios fetches and transcodes it server-side.

  3. 3

    Wait for ready

    Block on transcoding with --wait, or react to the video.ready webhook.

  4. 4

    Embed

    Hand the returned embed/playbackUrl to any player or iframe.

Fastest path

One command (or one MCP tool call) gets you started:

shell 
# an agent turns a generated clip URL into a streamable embed in one step
npx @rehelios/cli import https://model-output/clip.mp4 --wait

Agents: see /agents and /mcp.