voice-generator console

loading…

음성 합성 (Synthesize)

고급 설정

기록

목소리 등록 (Voice cloning reference)

3~10초 길이의 WAV 파일과 그 전사(transcript)를 올리면 재사용 가능한 목소리가 됩니다.

등록된 목소리

이름길이전사

API 사용 예

curl -X POST http://HOST:8080/api/v1/synthesize \
  -H 'Content-Type: application/json' \
  -d '{"text":"Hello from the API.","language":"English","instruct":"male, british accent"}' \
  -o out.wav

# code streaming (~2 kbps): decode on the client with omnivoice-decoder
omnivoice-decoder synth --server http://HOST:8080 --decoder ./decoder-weights \
  --text "A long text ..." --output out.wav

# streaming: raw 16-bit PCM arrives chunk by chunk
curl -N -X POST http://HOST:8080/api/v1/synthesize/stream \
  -H 'Content-Type: application/json' \
  -d '{"text":"A long text ...","options":{"first_chunk_duration":3}}' \
  | ffplay -f s16le -ar 24000 -ac 1 -

curl -X POST http://HOST:8080/api/v1/voices \
  -F name=alice -F ref_text='Reference transcript.' -F file=@ref.wav

grpcurl -plaintext -proto speech.proto -d '{"text":"Hi"}' HOST:50051 omnivoice.v1.SpeechService/Synthesize