Generate Speech API - Maya Research Maya1

This endpoint generates high-quality speech audio using pre-designed character voices. Supports streaming, emotional tags, and verbose output.

X-API-Key

string

required

API key for authentication

Request Body

voice_id

string

default:"Ava"

required

The unique identifier of a pre-designed voice from the Get Characters endpointRealistic Voices:

Ava
Chloe
Liam
Noah
James
Emma
Sophie
Oliver

Creative Character Voices:

AnimatedCartoon
Anime
Flirty
Seductively
AIMachineVoice
Cyborg
AlienSciFi
Pirate
Gangster
DarkVillain
Demon

text

string

required

The text to convert to speech.Max Length: 5,000 characters

stream

boolean

default:false

Enable streaming audio generation

true - Stream audio chunks as they’re generated (lower latency)
false - Return complete audio file after processing

curl --location 'https://v3.mayaresearch.ai/v1/tts/generate' \
--header 'Content-Type: application/json' \
--header 'X-API-Key: maya_YOUR_API_KEY_HERE' \
--data '{
  "voice_id": "Ava",
  "text": "Welcome back to another episode of our podcast! Today we are diving into an absolutely fascinating topic.",
  "stream": false
}'

Response

audio

binary

The generated audio file in WAV or MP3 format

metadata

object

Additional metadata about the generation (only when verbose is enabled)

Show Metadata Object

duration

number

Audio duration in seconds

sample_rate

number

Sample rate of the generated audio

character_used

string

Character ID used for generation

processing_time

number

Time taken to generate the audio in seconds

text_length

number

Length of the input text in characters

emotional_tags_detected

array

List of emotional tags detected in the input text

Audio file will be returned in WAV or MP3 format

Example Use Cases

{
  "voice_id": "Ava",
  "text": "Welcome back to another episode of our podcast! Today we're diving into an absolutely fascinating topic that I know you're going to love.",
  "stream": false
}

{
  "voice_id": "Emma",
  "text": "Oh my gosh, you won't believe what happened today! It was absolutely incredible. This is exactly the kind of story I love to share with you all.",
  "stream": false
}

{
  "voice_id": "Noah",
  "text": "Welcome back to another episode! Today we have something really special planned. Have you ever wondered how AI voices actually work? Let's dive right in and explore this amazing topic together.",
  "stream": true
}

{
  "voice_id": "Liam",
  "text": "Hello and welcome to today's show. We have an incredible lineup for you today that you absolutely won't want to miss.",
  "stream": false
}

{
  "voice_id": "James",
  "text": "Thank you for tuning in to another episode. Today's conversation is going to be particularly interesting as we explore some cutting-edge developments in the field.",
  "stream": false
}

{
  "voice_id": "Chloe",
  "text": "Good evening and welcome back. I'm absolutely delighted to have you with us today as we delve into this fascinating subject matter.",
  "stream": false
}

{
  "voice_id": "Sophie",
  "text": "Greetings everyone, and welcome to today's discussion. We have quite an extraordinary topic lined up that I'm confident you'll find thoroughly engaging.",
  "stream": false
}

{
  "voice_id": "Oliver",
  "text": "Welcome to this special episode where we explore some truly remarkable insights and perspectives on the subject at hand.",
  "stream": false
}

Streaming Response

When stream is set to true, the API returns audio chunks as they are generated for lower latency:

Streaming Example

const response = await fetch('https://v3.mayaresearch.ai/v1/tts/generate', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'X-API-Key': 'maya_YOUR_API_KEY_HERE'
  },
  body: JSON.stringify({
    voice_id: "Ava",
    text: "Welcome back to another episode of our podcast! Today we're diving into some truly fascinating content that I know you're going to absolutely love.",
    stream: true
  })
});

const reader = response.body.getReader();
const audioChunks = [];

while (true) {
  const { done, value } = await reader.read();
  if (done) break;
  audioChunks.push(value);
  
  // Process or play audio chunk immediately for real-time playback
  await playAudioChunk(value);
}

// Optionally combine all chunks for complete audio
const completeAudio = new Blob(audioChunks, { type: 'audio/wav' });

Use streaming mode for:

Real-time voice assistants and chatbots
Interactive applications where latency matters
Long-form content where you want to start playback immediately
Live customer support systems

Best Practices

Choose the Right Character

Browse available characters using the Get Characters endpoint and select one that matches your use case and target audience.

Match Voice to Content

Customer service → Friendly, patient, clear voices
Podcast → Conversational, warm, engaging voices
Audiobook → Clear, expressive, consistent voices
Professional → Authoritative, confident, polished voices

Optimize for Performance

Enable streaming for interactive applications
Cache character IDs for repeated use
Break long text into manageable chunks

Error Codes

Always implement proper error handling to gracefully manage API errors.

Code	Description	Resolution
400	Invalid request format	Check JSON syntax and ensure required fields are present
401	Authentication failed	Verify your API key is correct and properly formatted
403	Access denied	Check API permissions and usage quotas
404	Voice not found	Verify the voice_id exists using Get Characters endpoint
413	Text too long	Reduce text length to under 5,000 characters
500	Internal server error	Contact support if the error persists

Performance Tips

Cache Character IDs

Store and reuse character IDs for consistent voice across your application.

Enable Streaming

Streaming reduces perceived latency for interactive applications and long content.

Batch Requests

Generate multiple audio files in parallel when possible to maximize throughput.

Optimize Text

Break very long text into manageable chunks (under 1,000 characters) for better performance.

Pro tip: For consistent brand voice across your application, select the right voice once, test thoroughly, then reuse the same voice_id for all subsequent generations.

Authorizations

X-API-Key

string

header

required

API key for Maya1 API authentication

Body

application/json

voice_id

enum<string>

required

The unique identifier of a pre-designed voice

Available options:

Ava,

Chloe,

Liam,

Noah,

James,

Emma,

Sophie,

Oliver,

AnimatedCartoon,

Anime,

Flirty,

Seductively,

AIMachineVoice,

Cyborg,

AlienSciFi,

Pirate,

Gangster,

DarkVillain,

Demon

Example:

"Ava"

text

string

required

The text to convert to speech.

Maximum string length: 5000

Example:

"Welcome back to another episode of our podcast! Today we're diving into an absolutely fascinating topic that I know you're going to love. We've got some incredible insights to share with you, so let's get started right away."

stream

boolean

default:false

Enable streaming audio generation

Example:

false

Response

Successfully generated audio

Generated audio file in WAV format

Getting Started

Maya1 API Reference

Generate Speech with Maya1

Request Body

Response

Example Use Cases

Streaming Response

Best Practices

Error Codes

Performance Tips

Cache Character IDs

Enable Streaming

Batch Requests

Optimize Text

Authorizations

Body

Response

Getting Started

Maya1 API Reference

​Request Body

​Response

​Example Use Cases

​Streaming Response

​Best Practices

​Error Codes

​Performance Tips

Cache Character IDs

Enable Streaming

Batch Requests

Optimize Text

Authorizations

Body

Response

Request Body

Response

Example Use Cases

Streaming Response

Best Practices

Error Codes

Performance Tips