Hugging Face's logo

Spaces:
rajiv-ramteke
/
Video-Note-Taker
Sleeping

App Files Community
Video-Note-Taker / docs /api_documentation.md
rajiv-ramteke's picture
rajiv-ramteke
initial commit
6b64d63
preview code
|
Raw
History Blame Contribute Delete
2.86 kB

Deep-Dive Video Note Taker — API Documentation

Base URL 

http://localhost:8000/api/v1

Endpoints

1. Upload & Process Video

POST /process/upload

Upload a video file and start the AI processing pipeline.

Form Data:

Field Type Required Description
file File ✅ Video/audio file (≤500 MB)
openai_key String ❌ OpenAI API key (overrides .env)

Response:

{
  "status": "success",
  "message": "Upload successful. Processing started.",
  "data": {
    "job_id": "550e8400-e29b-41d4-a716-446655440000",
    "filename": "my_lecture.mp4",
    "size_mb": 45.3
  }
}

2. Job Status

GET /status/{job_id}

Poll the processing status of a job.

Response:

{
  "status": "success",
  "data": {
    "job_id": "550e8400...",
    "status": "processing",
    "progress": 60,
    "message": "Summarizing chunks with LLM...",
    "filename": "my_lecture.mp4"
  }
}

Status values: queued → processing → complete / error

3. List All Jobs

GET /status/

List all processing jobs in the current session.

4. Get Notes

GET /notes/{job_id}

Retrieve the structured notes for a completed job.

Response:

{
  "status": "success",
  "data": {
    "job_id": "550e8400...",
    "filename": "my_lecture.mp4",
    "notes": {
      "final_notes": "## Summary\n...",
      "highlights": [...],
      "action_items": [...],
      "chapters": [...],
      "transcript_segments": [...]
    },
    "markdown": "# Video Notes: my_lecture.mp4\n..."
  }
}

5. Download Notes (Markdown)

GET /notes/{job_id}/download

Download notes as a .md file.

6. Download Notes (JSON)

GET /notes/{job_id}/download/json

Download full structured notes as a .json file.

7. RAG Semantic Search

POST /query/

Search the video content using natural language.

Request Body:

{
  "job_id": "550e8400-e29b-41d4-a716-446655440000",
  "query": "What did they say about neural networks?",
  "top_k": 5
}

Response:

{
  "status": "success",
  "data": {
    "query": "What did they say about neural networks?",
    "results": [
      {
        "chunk_id": 3,
        "text": "Neural networks consist of layers of interconnected nodes...",
        "start_ts": "00:12:30",
        "end_ts": "00:14:00",
        "score": 0.87
      }
    ]
  }
}

8. Health Check

GET /health

{
  "status": "healthy",
  "app": "Deep-Dive Video Note Taker",
  "version": "1.0.0"
}

Interactive API Docs

When the server is running, visit: