EchoVox

A reproducible EchoVox text-to-speech API request

The EchoVox developer route accepts a voice ID, JSON synthesis options, and an x-api-key header, then returns audio with generation metadata in response headers.

Create and manage a key after sign-in, start with a short known script, and save the response to a file instead of printing binary audio in a terminal.

Use the production contract

Send a POST request to https://echovox-api.secondwindltd.xyz/api/v1/v1/text-to-speech/{voice_id}. Set Content-Type to application/json and pass the account key in the x-api-key header. The request body contains text, model_id, language, and stream. The current Web example uses model_id echo_turbo_v1.

Start with a controlled test

Use a known system voice such as system_maia and the sentence “Hello from EchoVox.” with language en. Save the body as echovox.mp3. This makes the result audible and repeatable while keeping the first request small enough to diagnose before a longer production script is introduced.

Record the response evidence

Successful audio responses include X-EchoVox-Generation-Id, X-EchoVox-Cache, and X-EchoVox-Seconds-Cost headers. Keep the generation ID, voice ID, exact source text, language, model, and output filename together. A cache HIT or MISS describes generation reuse; it is not an editorial approval status.

Handle failures and keys deliberately

A 401 response usually means the key is missing, invalid, or revoked. Validate voice IDs before large jobs, bound request timeouts, and never place a live key in browser source, screenshots, public repositories, or a public support message. Revoke a key when its integration or owner changes.

Production checklist

  • Create the key in an authenticated account
  • Use the production HTTPS endpoint
  • Save binary audio to a file
  • Record generation response headers
  • Review pronunciation and content before publishing
  • Revoke exposed or retired keys