Content and Uploads

Step 1: Create content

Create a content item (optionally tied to a channel).

GraphQL:

mutation {
  createContent(
    input: {
      hints: [{ type: FEATURE, containsAds: NO }]
      metadata: {
        title: "My Video"
        summary: "Video description"
      }
    }
  ) {
    id
    metadata { title }
  }
}

MCP: Use the create_content tool with title, summary, and type (e.g. FEATURE). The response includes the content id (DID).

Save the returned id (e.g. did:rad.live:content/feature/<id>); you need it for the next steps.

Step 2: Create an ingestion session (TUS)

Request a resumable upload session for the content. You get an endpoint to POST the file to.

GraphQL:

mutation {
  createContentAsset(
    id: "did:rad.live:content/feature/<id>"
    input: {
      filename: "my-video.mp4"
      size: 12345678
      video: { hint: NONE }
    }
  ) {
    id
    endpoint
    metadata { filename originalFilename }
  }
}

MCP: Use create_content_asset with contentId (the content DID), filename, size, and optional hint (e.g. NONE, PORTRAIT, EQUIRECTANGULAR).

The response includes endpoint (e.g. https://api.rad.live/ingestion/<session-id> or your TUS upload URL).

Step 3: Upload the file

POST the file to the endpoint from step 2:

• Headers: Authorization: Bearer <api-key>, Content-Type: multipart/form-data
• Body: The file (multipart/form-data)

Resumable uploads use the TUS protocol; the ingestion endpoint may point to a TUS server. Follow the session instructions returned by createContentAsset / create_content_asset.

Step 4: Submit for processing

After the file is uploaded, submit the content for transcoding. Use the same filename you passed to createContentAsset as assetName.

GraphQL:

mutation {
  submitContentForProcessing(
    id: "did:rad.live:content/feature/<id>"
    input: {
      assetName: "my-video.mp4"
      layout: LANDSCAPE
      encoder: RAD_ENCODER
    }
  ) {
    content { id }
    job { id status progress }
  }
}

MCP: Use submit_content_for_processing with contentId, assetName, and optional layout, encoder, enhance.

You get a TranscodeJob with id, status, and progress. Poll the content’s transcodeJob (or the job id) until the job is complete.

Step 5: Output assets

When transcoding is complete, the content’s outputAssets field is populated. Query it via GraphQL:

query {
  me {
    channel {
      features(limit: 10, start: 0) {
        id
        metadata { title }
        outputAssets {
          encoderType
          video { hls download preview }
          audio { download }
          images {
            thumbnails { urlFormat count }
            videoThumbnails { urlFormat count }
          }
          subtitles { srt vtt }
        }
      }
    }
  }
}

• video — hls (streaming), download (MP4), preview (may be null for some encoders)
• audio — download
• images — Thumbnails are described by ThumbnailSet: urlFormat (string with placeholder) and count. Replace the placeholder in urlFormat with zero-padded indices: XXXXXXX = 7-digit 0-based (MediaConvert); XX = 2-digit 1-based (Rad Encoder).
• subtitles — srt and vtt URLs (null for MediaConvert; present for Rad Encoder).

See Schema reference — Output assets and thumbnails for details.

Step 6: Publish (optional)

To make the content public:

GraphQL: publishContent(id: "did:rad.live:content/feature/<id>")
MCP: publish_content with id (content DID).

To revert to protected: unpublishContent / unpublish_content.