Skip to navigation

Kotoba Technologies APIs

Low-latency speech APIs for real-time and async applications.

Kotoba APIs are in private alpha, available to selected customers. Request access.

Kotoba Technologies provides two families of speech APIs:

Realtime APIs — WebSocket-based services for streaming speech with sub-second latency:

  • ASR — Automatic Speech Recognition. Stream audio and receive transcription deltas live.
  • STS — Speech-to-Speech Translation. Translate spoken audio between languages in real time.
  • TTS — Text-to-Speech. Synthesize natural-sounding speech from text.

Each realtime API speaks JSON over a single WebSocket connection and is described by an AsyncAPI specification.

Transcription API — A REST API for finished audio files:

  • Transcription — POST an audio file and receive the transcript in the response. Simpler than WebSocket when low latency is not required. Private endpoints add an asynchronous job API for long files.

Quickstart

Install the Python SDK (Python 3.10+):

pip install kotoba-sdk

Transcribe a file via the REST API:

import kotoba
client = kotoba.KotobaClient() # reads KOTOBA_API_KEY + KOTOBA_*_URL from env
result = client.asr.transcribe("clip.mp3", language="ja")
print(result.text)

Streaming TTS:

import kotoba
client = kotoba.KotobaClient()
with client.tts.stream(language="ja") as session:
session.synthesize("こんにちは。本日はよろしくお願いします。")
for event in session:
if event.type == "audio_chunk":
# event.audio is float32 PCM @ 24 kHz; pipe to a speaker
handle(event.audio)
elif event.type == "done":
break

Each capability has its own Python SDK page with end-to-end snippets: Speech to Text · Speech to Speech Translation · Text to Speech.

What’s next