A whole season

One project per season (or per language set), every episode in its input folder. Subfolders are kept all the way to the outputs.

season-1/                         output/
├── ep01.mp4                      ├── season 1/ep01.es.mp4
├── ep02.mp4          ──────▶     ├── season 1/ep02.es.mp4
└── extras/                       └── season 1/extras/
    └── making-of.mp4                 └── making-of.es.mp4

1. Create the project, stopped

autoStart: false lets you upload everything first. The service then sees the whole season in its first scan and works on it at full parallelism.

PROJECT=$(curl -s -X POST $API/projects -H "$AUTH" -H 'Content-Type: application/yaml' --data-binary @- <<'YAML' | jq -r .data.id
name: Fishing Channel — Season 1 — es
config:
  sourceMedia: { '@type': S3FolderSource }
  transcription: { '@type': Speechmatics, language: en }
  translations:
    - language: es
      engine: { '@type': Gemini }
      ttsOptions:
        provider: { '@type': Deepdub, voice: auto }
  outputs:
    - '@type': S3FolderOutput
      items:
        - '@type': VideoFile
          ext: mp4
          nameTemplate: '${name}.es.${ext}'
          useOriginalTracks: true
          dubTracks:
            - { language: es, title: Spanish }
options: { autoStart: false }
YAML
)

2. Upload the season

AWS CLI
API

With the upload role set up as the syncwords-vod profile:

INPUT=$(curl -s -H "$AUTH" $API/projects/$PROJECT | jq -r .data.storage.inputPath)
aws s3 sync ./season-1 "${INPUT}season 1/" --profile syncwords-vod \
  --exclude '*' --include '*.mp4' --include '*.mov'

With upload.mjs, keeping the folder structure in the names:

cd season-1
find . -type f \( -name '*.mp4' -o -name '*.mov' \) | sed 's|^\./||' | while read -r f; do
  node ../upload.mjs "$PROJECT" "$f" "season 1/$f"
done

3. Start

curl -s -X POST -H "$AUTH" $API/projects/$PROJECT/start | jq .data.status   # "ACTIVE"

4. Follow progress

curl -s -H "$AUTH" $API/projects/$PROJECT/files \
  | jq '.data | group_by(.status) | map({(.[0].status): length}) | add'

{ "done": 14, "pending": 7, "processing": 1 }

Wait for the season, and list what failed:

until curl -s -H "$AUTH" $API/projects/$PROJECT/files \
  | jq -e 'all(.data[]; .status == "done" or .status == "failed")' > /dev/null; do sleep 60; done

curl -s -H "$AUTH" $API/projects/$PROJECT/files | jq -r '.data[] | select(.status == "failed") | .name'

Retry each failed file (its name URL-encoded), then start the project again if it stopped:

curl -s -H "$AUTH" $API/projects/$PROJECT/files | jq -r '.data[] | select(.status == "failed") | .name | @uri' \
  | while read -r name; do curl -s -X POST -H "$AUTH" "$API/projects/$PROJECT/files/$name/retry"; done
curl -s -X POST -H "$AUTH" $API/projects/$PROJECT/start > /dev/null

5. Collect the outputs

AWS CLI
API

OUTPUT=$(curl -s -H "$AUTH" $API/projects/$PROJECT | jq -r .data.storage.outputPath)
aws s3 sync "$OUTPUT" ./season-1-es --profile syncwords-vod

curl -s -H "$AUTH" $API/projects/$PROJECT/outputs | jq -r '.data[].path' | while read -r path; do
  mkdir -p "season-1-es/$(dirname "$path")"
  url=$(curl -s -G -H "$AUTH" "$API/projects/$PROJECT/outputs/download" --data-urlencode "path=$path" | jq -r .data.url)
  curl -s -o "season-1-es/$path" "$url"
done

Tips for large batches

Distinct base names. ep01.mp4 and ep01.mov in one folder would write the same output.
Add episodes later by uploading into the same project and calling start. Files already done are not processed again.
Another language later? Create a second project for it and upload the season there too; changing the first project’s config would not reprocess its done files.
Throughput. Several files are processed at once, and the next file starts as soon as one finishes, so you can upload while earlier episodes are processing.

Get started

Concepts

Configure a project

Upload media

Examples

Going to production

Help

1. Create the project, stopped

2. Upload the season

3. Start

4. Follow progress

5. Collect the outputs

Tips for large batches

​1. Create the project, stopped

​2. Upload the season

​3. Start

​4. Follow progress

​5. Collect the outputs

​Tips for large batches

1. Create the project, stopped

2. Upload the season

3. Start

4. Follow progress

5. Collect the outputs

Tips for large batches