Frequently asked questions
How do I get an API key?
Sign up on the /get-api-key/ page using Google or an email and password. Once registered, your dashboard displays your unique API key, current balance, and usage statistics. You can also issue a new key at any time, which immediately invalidates the previous one.
What is the base URL?
Your API base URL is https://api.66mmr.info/v1. You send POST requests to https://api.66mmr.info/v1/audio/speech for synthesis. You can also use the official OpenAI SDKs by simply changing their base_url configuration to point to this address.
Can I use Python SDK?
Yes, the official OpenAI Python SDK works perfectly. You just need to update the base_url to https://api.66mmr.info/v1 and provide your Auralux API key. No custom SDK is required; standard HTTP requests or the OpenAI library are sufficient for integration.
How do I handle long texts?
Split your input text into chunks of up to 4096 characters. Send each chunk as a separate POST request to the speech endpoint. Then, join the resulting audio files in your application code to create a single continuous audio stream for the user.
What is the character limit?
Each API request accepts up to 4096 characters in the input field. Spaces and punctuation count toward this limit. If your text exceeds this length, you must split it into multiple requests and concatenate the returned audio files manually.
Do I need SSML?
No, SSML is not supported. The API accepts plain text input. You control pacing and emphasis by adjusting the speed parameter between 0.25 and 4.0. For more granular control, split your text into smaller segments and adjust speeds individually.
How do I set the language?
Include the optional 'language' field in your JSON body with an ISO code. If omitted, the system guesses the language from the script, defaulting to English for Latin text. Explicitly setting it ensures accurate pronunciation for non-English Latin-script inputs.
What happens on rate limit?
You are limited to 60 requests per minute per key (10 before top-up). Requests are processed sequentially, so high load may cause queuing delays. Failed requests are free, but successful ones count against your limit and balance immediately.
Is there a WebSocket endpoint?
No, there is no WebSocket endpoint or real-time streaming support. You send a text payload and receive a complete audio file in the response. This synchronous model simplifies integration for voice assistants that do not require partial audio chunks.
How do I rotate keys?
Generate a new key from your dashboard at any time. The new key becomes active immediately, and the old one stops working at once. This allows seamless rotation without downtime. Each account maintains only one active key at a time.
Can I use multiple keys?
No, each account has exactly one active API key. If you generate a new key, the previous one is instantly invalidated. You cannot use multiple keys simultaneously for a single account; you must rotate them sequentially.
What is the response format?
The response contains the generated audio file directly in the response body. You can specify the format via the 'response_format' field: mp3, wav, opus, flac, or pcm. The audio is returned in the response body, ready for immediate playback or storage in your application.
How do I detect errors?
Check the HTTP status code of the response. Failed synthesis requests are free, meaning no charge is applied if the API returns an error. Inspect the response body for error details to debug issues like invalid parameters or rate limits.
Is streaming supported?
No, streaming is not supported. The API returns the full audio file after synthesis completes. For voice assistants, you must wait for the complete file before playback.
First request
curl
curl https://api.66mmr.info/v1/audio/speech \
-H "Authorization: Bearer $TTS_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"tts-1","voice":"aria","input":"Hello from the API.","response_format":"mp3"}' \
--output hello.mp3Python (openai)
from openai import OpenAI
client = OpenAI(base_url="https://api.66mmr.info/v1", api_key="tts-...")
with client.audio.speech.with_streaming_response.create(
model="tts-1", voice="nova", input="Hello from Python.", response_format="mp3"
) as response:
response.stream_to_file("hello.mp3")Node.js (openai)
import OpenAI from "openai";
import fs from "node:fs";
const client = new OpenAI({ baseURL: "https://api.66mmr.info/v1", apiKey: process.env.TTS_API_KEY });
const res = await client.audio.speech.create({ model: "tts-1", voice: "milo", input: "Hello from Node.js." });
fs.writeFileSync("hello.mp3", Buffer.from(await res.arrayBuffer()));| voice | |
|---|---|
aria | female |
nova | female |
luna | female |
sage | female |
felix | male |
milo | male |
atlas | male |
orion | male |
| language | |
|---|---|
ar | العربية |
zh | 中文 |
da | Dansk |
nl | Nederlands |
en | English |
fi | Suomi |
fr | Français |
de | Deutsch |
el | Ελληνικά |
he | עברית |
hi | हिन्दी |
it | Italiano |
ja | 日本語 |
ko | 한국어 |
ms | Bahasa Melayu |
no | Norsk |
pl | Polski |
pt | Português |
ru | Русский |
es | Español |
sw | Kiswahili |
sv | Svenska |
tr | Türkçe |
Authentication
Your API key authenticates every request to the speech endpoint. Pass it in the HTTP header as Authorization: Bearer tts-.... This simple token ensures only your account’s balance is charged for character usage.
You can manage keys directly from the dashboard. The system supports instant rotation, allowing you to switch tokens without interrupting active integrations or requiring downtime in your voice assistant workflow.
Bearer token format
Include the header Authorization: Bearer tts-your-key-here in every POST request to /v1/audio/speech. The key must remain valid and associated with an account that has sufficient balance.
Using the OpenAI SDK or standard HTTP libraries, this header format is all that is required to identify your application.
- Format:
Bearer tts-<your-key> - Case-sensitive: ensure exact character match
- Included in every request header
Single active key
Each account maintains one active API key at any given time. This simplifies security auditing for development teams, as there is only one credential to track in your environment variables.
If you suspect a leak or need to rotate credentials, generating a new key immediately invalidates the previous one.
- One key per account
- Simplified security tracking
- No complex permission scopes
Instant key rotation
When you issue a new key, the old one stops working immediately. This ensures that any compromised credentials lose access instantly, protecting your character balance from unauthorized usage.
You can generate a new key at any time via the dashboard without waiting for a cooldown period.
- Old key invalidates immediately
- No waiting period required
- Secure credential management
Endpoints
Three core endpoints handle your voice assistant’s needs. The speech endpoint generates audio from text, while the voices and models endpoints provide reference data for integration.
Use the voices endpoint to retrieve available presets and supported languages. The models endpoint confirms the available engine versions without generating audio.
POST /v1/audio/speech
Send text and receive an audio file. This is your primary endpoint for generating speech in your application.
Supports standard request fields like model, input, voice, response_format, and speed. Text is limited to 4096 characters per request.
- Endpoint: POST /v1/audio/speech
- Returns: Audio file based on
response_format - Input: JSON body with
input(text),voice, andmodel
GET /v1/voices
Retrieve the list of eight available voice presets and their details. Use this to let users select a voice or to configure your bot’s default speaker.
Returns supported languages, formats, and price information for each voice configuration.
- Endpoint: GET /v1/voices
- Returns: List of voices, languages, formats, and pricing
GET /v1/models
Check the available model versions. Currently, this confirms support for tts-1 and tts-1-hd, which use the same engine and pricing.
Use this endpoint to verify model availability before sending synthesis requests.
- Endpoint: GET /v1/models
- Returns: Available model IDs (e.g., tts-1, tts-1-hd)
Request parameters
Configure your synthesis requests with precise parameters to ensure optimal audio quality and performance. The API supports two model variants for different latency and fidelity needs, allowing you to balance cost and quality based on your application's requirements.
Model selection (tts-1)
Choose between "tts-1" for standard generation or "tts-1-hd" for high-definition audio. Both models use the same underlying engine and pricing structure, so you can switch based on your specific quality needs without changing your cost model.
Input text limit
Each request accepts up to 4096 characters of input text. Spaces and punctuation count toward this limit. For longer scripts, split your text into chunks, generate each part, and join the resulting audio files to create a continuous track.
Voice and language codes
Select from eight preset voices, such as "aria" for warm narration or "atlas" for neutral news. You can optionally specify an ISO language code (e.g., "en", "de", "ja"). If omitted, the system guesses the language, defaulting to English for Latin-script text.
- Supported voices: aria, nova, luna, sage, felix, milo, atlas, orion
- Supported formats: mp3, wav, opus, flac, pcm
- Speed range: 0.25 to 4.0
Response formats
The API returns audio files in the format you specify, allowing you to choose the best balance between file size and quality for your voice assistant’s playback requirements. Every generated file carries an inaudible watermark to mark it as synthetic speech, ensuring clear attribution without affecting audio fidelity.
MP3 output
MP3 is the default and most widely supported format. It offers excellent compression for efficient network transfer.
- Default response format
- Optimized for network efficiency
- Universal playback support
WAV output
Choose WAV for uncompressed, high-quality audio. This format is suitable for local processing or when raw audio data is required.
- Uncompressed audio
- Highest fidelity
- Ideal for local processing
OPUS and FLAC support
OPUS and FLAC are available as response formats alongside mp3, wav, and pcm. Both formats give you flexibility depending on whether you prioritize bandwidth savings or perfect audio preservation for your chatbot’s voice layer.
- OPUS: efficient streaming
- FLAC: lossless quality
- Flexible archival options
OpenAI SDK integration
The API mirrors the OpenAI speech endpoint shape, allowing you to reuse existing codebases with minimal changes. By adjusting the base URL, your current integration logic works immediately without rewriting request handlers or parsing logic.
This compatibility means your team can adopt Auralux without retraining engineers on a new SDK. You keep your preferred language bindings and simply point them to our infrastructure.
Change base_url
Update your client configuration to point to https://api.66mmr.info/v1. This single change redirects all requests to our TTS engine while maintaining the standard request structure you already know.
Ensure your authentication header remains formatted as Bearer tts-<your-key> to validate requests correctly against our endpoint.
- Base URL: https://api.66mmr.info/v1
- Standard JSON body format
- Compatible with OpenAI SDK clients
Use official SDK
Leverage the official OpenAI SDK for Python, Node.js, or Go. The SDK handles serialization and response parsing exactly as it does for OpenAI, delivering the audio file directly to your application.
This approach simplifies development for voice assistants and chatbots, as you can test with familiar tools before switching providers.
- Works with OpenAI Python/Node SDKs
- Standard response parsing
- No custom client needed
Swap credentials easily
Replace your OpenAI key with your Auralux key in the same configuration variable. The SDK treats both keys as standard bearer tokens, ensuring a seamless transition.
Key rotation is instant; issuing a new key invalidates the old one immediately, so your SDK updates reflect changes in real-time without downtime.
- Standard Bearer token auth
- Instant key rotation
- Zero config changes needed

Rate limits
Your API key allows 60 requests per minute once you top up your account. Before your first top-up, you are limited to 10 requests per minute. The limit increases to 60/minute after your first top-up via USDT on the TRON network.
60 requests per minute
After your first top-up via USDT on the TRON network, your rate limit increases to 60 requests per minute. This allows your voice assistant to handle moderate conversation flows without hitting a ceiling.
10 requests initially
New accounts start with a limit of 10 requests per minute. This is sufficient for testing and development before you process larger batches of user interactions.
Queue processing
Requests are processed sequentially. If you exceed your limit, further requests wait in a queue rather than failing immediately. This ensures fair usage but may add slight latency during high load.
Error handling
When integrating Auralux into your voice assistant workflow, understanding HTTP status codes helps you handle failures gracefully. Our API returns standard codes to indicate authentication issues, rate limits, or internal processing errors. Proper error handling ensures your application remains responsive even when the service experiences temporary load or when your API key is rotated.
Most errors are transient or indicate a clear configuration issue. By checking the response body, you can determine whether to retry immediately, wait for the rate limit window to clear, or rotate your key.
401 Unauthorized
This error occurs when the Authorization header is missing or contains an invalid token. It can also happen if you have issued a new key and the old one has been revoked. Check your key in the dashboard or rotate it via the API.
Ensure your key starts with tts- and is sent as a Bearer token. If you recently rotated your key, the previous one stops immediately.
- Check for typos in the Bearer token.
- Verify the key is active in your dashboard.
- Rotate the key if the old one was invalidated.
429 Too Many Requests
You have exceeded the rate limit of 60 requests per minute per key. New keys start with a limit of 10 requests per minute. The response includes a Retry-After header indicating how many seconds to wait before retrying.
If you are processing long texts, remember that requests are queued. High load may increase wait times, but the rate limit applies to submitted requests, not processing duration.
- Implement exponential backoff for retries.
- Monitor the
Retry-Afterheader value. - Split long texts into chunks under 4096 characters to reduce queue wait times.
500 Internal Error
An unexpected error occurred on the server side during synthesis. This is rare but can happen during peak load. Failed requests are free, so you do not lose credits for these errors.
Wait a few seconds and retry the request. If the error persists, check the /v1/models endpoint to ensure the model is available.
- Retry after a short delay.
- Verify model availability via GET /v1/models.
- Contact support if the issue persists.
Curl example
To get started, send a POST request to the speech endpoint with your Bearer token. The API returns the audio file directly, allowing you to integrate it into your voice assistant's playback logic immediately.
Basic request structure
Use a standard HTTP client like cURL to send your request. Ensure you include the Authorization: Bearer tts-... header with your active key.
- Endpoint:
https://api.66mmr.info/v1/audio/speech - Method:
POST - Header:
Authorization: Bearer YOUR_KEY
JSON body format
Provide the text, voice, and model in the JSON body. Text is limited to 4,096 characters per request.
model: "tts-1" or "tts-1-hd"input: "Your text here"voice: "nova"response_format: "mp3"
Audio file retrieval
The response contains the audio binary. Save it to a temporary file or stream it directly to your user's device. For longer texts, split the input, process each part, and join the resulting audio files in your application code to create a single continuous audio file for the user.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. This model simplifies budgeting for large teams, as expenses are tied strictly to character count rather than unpredictable minute-based tiers.
23 languages
Any voice speaks any of these languages.
