API Documentation

Getting Started

FloweryAI provides a simple REST API for converting speech to text. Our API endpoints accept audio files and return accurate transcriptions.

Base URL:

https://floweryai.l5.ca

Available Models

Model Name	Description
whisper-tiny	Lightweight model - fastest model
whisper-base	Base model - good balance of speed and accuracy

Endpoints

GET /stt/models

List all available speech-to-text models

Response:

{
  "available_models": [
    {
      "name": "whisper-tiny",
      "description": "Tiny model for Whisper speech-to-text",
      "provider": "floweryai",
      "input_modalities": ["audio"],
      "output_modalities": ["text"]
    },
    {
      "name": "whisper-base",
      "description": "Base model for Whisper speech-to-text",
      "provider": "floweryai",
      "input_modalities": ["audio"],
      "output_modalities": ["text"]
    }
  ]
}

Example:

curl https://floweryai.l5.ca/stt/models

POST /stt/transcribe

Transcribe an audio file to text

Request:

Form data with:

audio - The audio file to transcribe (WAV, MP3, M4A, FLAC)
model - (Optional) Model to use. Default: whisper-tiny
timestamps - (Optional) Set to "true" to include timestamps. Default: false

Response:

{
  "transcription": "This is the transcribed text from the audio file."
}

Response with timestamps:

{
  "transcription": "[00:00:00.000 --> 00:00:05.520] This is the transcribed text from the audio file.\n[00:00:05.520 --> 00:00:10.560] With timestamps showing when each segment was spoken."
}

Example:

curl -X POST https://floweryai.l5.ca/stt/transcribe \
  -F "audio=@recording.wav" \
  -F "model=whisper-tiny" \
  -F "timestamps=true"

Code Examples

// Example: Transcribe an audio file
async function transcribeAudio(audioFile, model = 'whisper-tiny', timestamps = false) {
  const formData = new FormData();
  formData.append('audio', audioFile);
  formData.append('model', model);
  formData.append('timestamps', timestamps ? 'true' : 'false');
  
  const response = await fetch('https://floweryai.l5.ca/stt/transcribe', {
    method: 'POST',
    body: formData
  });

  return await response.json();
}

// Usage with file input
document.getElementById('audioInput').addEventListener('change', async (e) => {
  const file = e.target.files[0];
  if (file) {
    try {
      // Get transcription with timestamps
      const result = await transcribeAudio(file, 'whisper-tiny', true);
      console.log('Transcription with timestamps:', result.transcription);
    } catch (error) {
      console.error('Error:', error);
    }
  }
});

import requests

def transcribe_audio(audio_file_path, model='whisper-tiny', timestamps=False):
    """
    Transcribe an audio file using FloweryAI API
    
    Parameters:
        audio_file_path (str): Path to the audio file
        model (str): Model to use for transcription
        timestamps (bool): Include timestamps in the transcription
        
    Returns:
        str: Transcribed text
    """
    url = 'https://floweryai.l5.ca/stt/transcribe'
    
    with open(audio_file_path, 'rb') as audio_file:
        files = {'audio': audio_file}
        data = {
            'model': model,
            'timestamps': 'true' if timestamps else 'false'
        }
        
        response = requests.post(url, files=files, data=data)
        
    if response.status_code == 200:
        return response.json()['transcription']
    else:
        raise Exception(f"Error: {response.status_code}, {response.text}")

# Usage
try:
    // Get transcription with timestamps
    transcription = transcribe_audio('recording.wav', model='whisper-tiny', timestamps=True)
    print(f"Transcription with timestamps: {transcription}")
except Exception as e:
    print(f"Failed to transcribe: {e}")

# List available models
curl https://floweryai.l5.ca/stt/models

# Transcribe an audio file using the default model (whisper-tiny)
curl -X POST https://floweryai.l5.ca/stt/transcribe \
  -F "audio=@recording.wav"

# Transcribe with a specific model
curl -X POST https://floweryai.l5.ca/stt/transcribe \
  -F "audio=@recording.wav" \
  -F "model=whisper-tiny"

# Transcribe with timestamps
curl -X POST https://floweryai.l5.ca/stt/transcribe \
  -F "audio=@recording.wav" \
  -F "timestamps=true"