Create from PPTX/PPT

post

https://apis.elai.io/api/v1/videos/from-presentation

Create a new video from ppt/pptx presentation

Recent Requests

Time	Status	User Agent
Retrieving recent requests…

Loading…

🚧
This API accepts only multipart/form-data with file
We don't support JSON base64-encoded strings. File should be sent with key "file" as a part of multipart/form-data request.

We accept only PPT/PPTX format.

Note: If request will be too long, you will recieve an http 206 status code. That means that conversion will take longer time than usual. Converted video will appear in the List videos request.

Helpful links

Send multipart/from-data in Python

Send multipart/from-data with NodeJS/Axios

Body Params

file

required

The .pptx or .pdf to convert. Required — the request fails with File is missing if the form has no part named file.

mode

enum

Conversion mode. May be repeated in the form to combine modes, so it is either one value or several.

ignoreFonts — do not try to reproduce the presentation's fonts
ignoreReadOnlyFonts — reproduce fonts except those the file marks read-only
static — convert slides as flat images rather than editable objects

folderId

string | null

Folder to create the video in. Omit or send null for the root folder.

templateId

string

Template whose slide layouts the converted slides adopt. See GET /templates.

brandKitId

string

Brand kit to apply during conversion. Validated against your workspace — an unknown id fails the request.

gender

string

Preferred avatar gender for the generated slides. Free-form string, not an enum.

avatarType

string

Preferred avatar type for the generated slides. Free-form string, not an enum.

fontReplacements

string

Font substitutions to apply during conversion.

llmEnhancements

string

JSON string (not an object — this is a multipart form field) describing LLM work to do during conversion. Today only generate-missing-speech is consumed:

{"generate-missing-speech": {"instructions": "Keep it under 20 words per slide"}}

Parsed leniently: malformed JSON or an unexpected shape is dropped and the upload still succeeds, so a typo here fails silently rather than returning 400.

quizSettings

string

JSON string (not an object) describing how quiz generation should behave. Parsed strictly — malformed JSON or an unknown key returns 400:

{"quizType": "single", "placement": "end", "count": 3, "answersCount": 4, "feedback": true}

quizType: single, multiple, true-false, mixed. placement: distributed, end. count: integer ≥ 0. answersCount: integer 2–6. feedback: boolean — whether a viewer sees an explanation after answering.

logId

string

Your own correlation id, echoed into our conversion logs. Generated for you if omitted.

Responses

`GET /videos` for it rather than retrying, which would convert the file a second time.

`AppError` defaults to 400, so this endpoint does not return a 500.

🚧This API accepts only multipart/form-data with file

Helpful links

200Conversion finished. The body is the new video id as a bare JSON string, not an object.

206Conversion is still running. It exceeded the request timeout of just under 5 minutes, so the response returned early — this is not a failure. The video is still being built: poll GET /videos for it rather than retrying, which would convert the file a second time.

400The upload could not be converted. Both conversion failures are 400s — AppError defaults to 400, so this endpoint does not return a 500.

🚧
This API accepts only multipart/form-data with file

`GET /videos` for it rather than retrying, which would convert the file a second time.

`AppError` defaults to 400, so this endpoint does not return a 500.