Voice cloning inference on A100 GPUs.
$5/mo for 50,000 characters.
The local tool handles cloning on your hardware. The cloud API handles synthesis from anywhere, without Apple Silicon.
Use afterwords to create a voice profile from a 15-second clip — runs locally on Apple Silicon, nothing leaves your machine.
Upload your WAV reference file and a transcript to your cloud account. Voice is available instantly to all API calls.
POST /v1/synthesize from any language, any device. Returns a job ID — poll GET /v1/jobs/:id until ready, then download the WAV. No Apple Silicon required.
Flat-rate billing. No hidden fees. Overage charged at $0.002 per additional 1,000 characters.
Bearer token auth. Async synthesis with polling. Cache hits return immediately.
# 1. Submit a synthesis job
curl -X POST https://afterwords-api.adrianwedd.workers.dev/v1/synthesize \
-H "Authorization: Bearer aw_<your-key>" \
-H "Content-Type: application/json" \
-d '{"voice_id":"<uuid>","text":"Hello, this is my cloned voice."}'
# → 202 Accepted
# {"job_id":"550e8400-...","status":"pending"}
# 2. Poll for completion
curl https://afterwords-api.adrianwedd.workers.dev/v1/jobs/550e8400-... \
-H "Authorization: Bearer aw_<your-key>"
# → {"job_id":"550e8400-...","status":"ready","audio_url":"..."}
# 3. Download audio
curl https://afterwords-api.adrianwedd.workers.dev/v1/jobs/550e8400-.../audio \
-H "Authorization: Bearer aw_<your-key>" \
-o output.wavFull specification: API reference →