Skip to main content

​
Endpoint

POST /api/generate
Generates a complete video presentation with AI-generated content, narration, visuals, and animations.
This is a long-running operation that can take several minutes. Use the Progress endpoint to track generation status in real-time.

​
Request Body

​
topic
string
required
The topic or subject for the presentation. This will be used to generate content, images, and animations.Example: "Introduction to Machine Learning"
​
num_slides
integer
default:"5"
The number of slides to generate in the presentation.Default: 5
​
language
string
default:"english"
The language for voice narration.Default: "english"
​
tone
string
default:"formal"
The tone of the presentation narration.Default: "formal"

​
Response

​
status
string
Status of the generation request. Returns "success" when completed.
​
message
string
Human-readable message describing the result.
​
content_data
object
The generated presentation content structure including slides, titles, and content.Contains:
  • title: Presentation title
  • slides: Array of slide objects with titles, content, and visual specifications
​
script_data
object
The generated narration scripts with timestamps for each slide.Contains:
  • slide_scripts: Array of scripts with start_time, end_time, and narration_text
  • total_duration: Total video duration in seconds
​
video_path
string
Full file system path to the generated video file.
​
video_filename
string
The filename of the generated video. Use this with the Video endpoint to stream the video.

​
Generation Process

The endpoint performs the following steps:
  1. Content Generation (10%): Creates presentation structure and slide content
  2. Script Generation (20%): Generates narration scripts with timestamps
  3. Audio Generation (30-49%): Generates voice narration for each slide
  4. Visual Generation (50-80%): Creates slides with text, images, or animations
  5. Video Composition (85-95%): Combines slides and audio into final video
  6. Completion (100%): Returns video information
Track progress in real-time using the Progress endpoint with the generation ID derived from the sanitized topic name.

​
Example Request

curl -X POST http://localhost:8000/api/generate \
  -H "Content-Type: application/json" \
  -d '{
    "topic": "Introduction to Machine Learning",
    "num_slides": 5,
    "language": "english",
    "tone": "formal"
  }'

​
Example Response

{
  "status": "success",
  "message": "Presentation video generated successfully",
  "content_data": {
    "title": "Introduction to Machine Learning",
    "slides": [
      {
        "slide_number": 1,
        "title": "What is Machine Learning?",
        "content_text": "Machine learning is a subset of artificial intelligence...",
        "needs_animation": true,
        "needs_image": false,
        "duration": 12.5
      }
    ]
  },
  "script_data": {
    "slide_scripts": [
      {
        "slide_number": 1,
        "start_time": 0,
        "end_time": 12.5,
        "narration_text": "Welcome to our introduction to machine learning..."
      }
    ],
    "total_duration": 62.3
  },
  "video_path": "/path/to/output/Introduction_to_Machine_Learning_final.mp4",
  "video_filename": "Introduction_to_Machine_Learning_final.mp4"
}

​
Error Responses

​
500 Internal Server Error

Returned when video generation fails: 
{
  "detail": "Error: Failed to generate content - API key not configured"
}

​
Generation ID

The generation ID is derived from the topic by:
  1. Taking the first 30 characters
  2. Replacing spaces with underscores
  3. Removing special characters (:, /, \, ", ', ?, !)
Example: "Introduction to Machine Learning" becomes "Introduction_to_Machine_Learn" This ID is used to track progress via the SSE endpoint.

​
Visual Types

Slides can have three types of visuals (mutually exclusive):
  1. Animation: Generated using Manim for mathematical/technical concepts
  2. Image: Fetched from external sources based on keywords
  3. Text-only: Default slide with styled text content
Each slide will have only ONE visual type. If both animation and image are flagged, animation takes priority.

Build docs developers (and LLMs) love

Get started for freeTalk to us