Skip to main content

​
Overview

The GistMag Text-to-Speech API converts text into natural-sounding speech using KittenTTS (local, English) and Google Cloud Text-to-Speech (all other languages). It also provides Speech-to-Text transcription using the OpenAI Whisper API. The API supports multiple languages, multiple voices, streaming, background music, and high-quality audio transcription.

​
Features

Basic TTS

Convert text to speech with a simple API call

Streaming

Stream audio in real-time as it’s generated

Batch Processing

Process long text in batches with pauses

With Music

Generate speech with background music

Speech-to-Text

Transcribe audio files to text

Change Speed

Adjust playback speed of audio files

Add Music

Add background music to existing audio

Voices

Browse and select from KittenTTS and Google Cloud voices

Languages

List all supported languages

​
Supported Languages

The TTS API supports multiple languages, including:
  • English (en)
  • Spanish (es)
  • French (fr)
  • German (de)
  • Italian (it)
  • Portuguese (pt)
  • Japanese (ja)
  • Korean (ko)
  • Chinese (zh)

​
Audio Formats

  • Input: Text (plain string)
  • Output: WAV (uncompressed) or MP3 (compressed) audio files
  • Streaming: MP3 format for efficient streaming

​
Engines

​
Text-to-Speech

TTS runs on two engines, and requests are routed automatically — you never have to pick one: How routing works:
  1. Naming a voice selects its engine — the eight names above run on KittenTTS, anything else is a Google voice.
  2. With no voice named, English uses KittenTTS and other languages use Google.
  3. Setting pitch or volume_gain_db routes to Google, since KittenTTS has no equivalent.
You can override this with the optional provider field ("kitten" or "google") on /tts, /tts/stream, /tts/batch and /tts/with-music.
Selecting a KittenTTS voice while also setting pitch or volume_gain_db will ignore those two parameters — the named voice takes priority. Omit the voice name to have the request routed to Google instead.
GET /tts/providers reports which engines are currently available.

​
Speech-to-Text

The API uses OpenAI Whisper API for STT, which provides:
  • High-accuracy transcription with automatic language detection
  • Support for many audio formats (MP3, WAV, M4A, FLAC, OGG, etc.)
  • Multi-language support with automatic language detection
  • Robust handling of various accents and audio qualities

​
Quick Start

The response will be an audio file that you can download or play directly.