One API. All the Best Speech-to-Text Models.

curl -X POST "https://api.speechall.com/v1/transcribe?model=deepgram.nova-3" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  --data-binary "@/path/to/audio.wav"

# -- Use Speechall SDK --

transcript = speechallApi.transcribe(
    model=TranscriptionModelIdentifier.OPENAI_DOT_WHISPER_MINUS_1,
    body=f.read()
).text

# -- OR Use OpenAI SDK --

from openai import OpenAI
client = OpenAI(base_url="https://api.speechall.com/v1/openai-compatible")
audio_file = open("/path/to/audio.wav", "rb")

transcription = client.audio.transcriptions.create(
  file=audio_file,
  model="assemblyai.best", # You can use any model you want, just change the model names
  response_format="verbose_json",
  timestamp_granularities=["word", "segment"]
)

print(transcription.words)

// -- Use Speechall SDK --

const transcript = await speechApi.transcribeRemote({
  file_url: 'https://example.com/audio.mp3',
  model: 'openai.whisper-1' // Replace the model name, and you are good to go!
}).data.text;

// -- OR Use OpenAI SDK --

import { OpenAI } from "openai";
let openai = new OpenAI({ baseURL: 'https://api.speechall.com/v1/openai-compatible' });

const transcription = await openai.audio.transcriptions.create({
  file: fs.createReadStream("/path/to/audio.wav"),
  model: "assemblyai.best", // You can use any model you want, just change the model names
  response_format: "verbose_json",
  timestamp_granularities: ["word", "segment"]
});

console.log(transcription.words);