Build withJapanese AI

Made for developers building Japanese products. LLMs, voice, translation. Bring your own harness and build in minutes.

curl https://api.shisa.ai/openai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $SHISA_API_KEY" \
  -d '{
    "model": "shisa-ai/shisa-v2.1-llama3.3-70b",
    "messages": [{
      "role": "user",
      "content": "こんにちは、シーサーです！"
    }]
  }'

Shisa LLM

OpenAI-compatible chat completions, purpose-tuned for Japanese and bilingual JA/EN tasks. Drop-in with the OpenAI SDKs.

LLM docs

Speech Recognition

Accurate Japanese speech-to-text with automatic language detection, hotword boosting, configurable voice activity detection (VAD), and realtime streaming.

ASR docs

Text-to-Speech

Natural-sounding Japanese voices with streaming, multiple audio formats, and a catalogue of ready-to-use voice presets.

TTS docs

Translation

Japanese ↔ English translation with context awareness, streaming output, and an OpenAI-style response shape.

Translation docs

Get an API key

Every Shisa API uses the same key. Create one in the Shisa platform, then pass it as a bearer token — see Authentication for the shared bearer-token header convention. New accounts receive $10 in free API credits to spend across any service.

API host

All HTTPS and WSS services share a single host:

api.shisa.ai

Service	Transport / method	Path
LLM (chat completions)	`HTTPS POST`	`/openai/v1/chat/completions`
Speech Recognition (batch)	`HTTPS POST`	`/asr/srt/audio_llm`
Speech Recognition (realtime)	`WSS`	`/ws/asr/realtime`
Text-to-Speech (HTTP)	`HTTPS POST`	`/tts`
Text-to-Speech (streaming)	`WSS`	`/ws/tts/realtime`
Translation	`HTTPS POST`	`/translate/`

Ready to build? Start with the LLM quickstart.