Generate AI video API

Asynchronous API: generate AI video

Create a narrated video from plain text, a JSON video script, an uploaded file or an HTTPS URL. A JSON input may contain a scripts array; the service then creates one video for each script.

Endpoint: POST https://mxprocess.com/api/files/start-process
Authentication: Bearer JWT from POST /api/login_check
Execution: asynchronous.

Parameters

imageSource — wikimedia_commons or ai. With Wiki Commons, named people found in the text are used to find relevant images; when no named person is found, AI images are used.
language — narration language: en, fr, es, de, hi, zh or ar.
voiceGender — male or female.
watermarkText — optional text displayed in white in the top-right corner of every generated video (maximum 80 characters).

Start from text

curl --location --request POST 'https://mxprocess.com/api/files/start-process' \
--header 'Authorization: Bearer YOUR_TOKEN' \
--form 'actionName=GenerateAiVideo' \
--form 'fileContent=A bear runs through a forest.' \
--form 'parameters={"imageSource":"ai","language":"en","voiceGender":"female","watermarkText":"My YouTube channel"}' \
--form 'async=true'

Instead of fileContent, send file=@"/path/to/script.txt" or an HTTPS fileUrl. For structured input, send a JSON object containing a scripts array.

Track and download results

Use the returned trackingUrl with the same Bearer token. When the action is complete, actionResultsDownloadURLs contains the generated video files and any accompanying metadata.

curl --location 'https://mxprocess.com/api/actions/get-action-infos/ACTION_ID' \
--header 'Authorization: Bearer YOUR_TOKEN'

Open the service