2026-05-27·8 min read·By Petar

Text to Speech API Python Tutorial (2026): Audexum Quickstart Guide

Learn how to call a text-to-speech API in Python using the requests library. Full code examples for voice selection, WAV output, error handling, and pricing comparison.

Integrating text-to-speech into a Python project takes less than 20 lines of code. This guide walks through everything: authentication, choosing a voice, saving WAV output, handling errors, and picking the right plan for your usage.

All examples use the Audexum TTS API, which supports 43 voices across 33 languages and has a free tier with no credit card required.

Prerequisites

You only need the requests library. No SDK, no framework.

bash

pip install requests

Get a free API key at audexum.com/signup. Keys follow the format sk_live_abc123xyz. The free plan gives you 10,000 characters per month with no card required.

First API Call

The endpoint accepts a JSON body with text, voice_id, and optionally language. It returns raw WAV audio.

python

import requests

API_KEY = "sk_live_abc123xyz"
API_URL = "https://audexum.com/api/v1/tts"

response = requests.post(
    API_URL,
    headers={
        "Authorization": f"Bearer {API_KEY}",
        "Content-Type": "application/json",
    },
    json={
        "text": "Hello, this is a test of the Audexum text-to-speech API.",
        "voice_id": "en_us_female_01",
    },
)

if response.status_code == 200:
    with open("output.wav", "wb") as f:
        f.write(response.content)
    print("Saved output.wav")
else:
    print(f"Error {response.status_code}: {response.text}")

Run it and you get output.wav in the current directory. That is the complete working implementation.

Selecting a Voice

Audexum has 43 voices across American English, British English, Spanish, French, German, Japanese, Korean, Arabic, Hindi, Bulgarian, Italian, Portuguese, and more. List available voices with a GET request:

python

import requests

API_KEY = "sk_live_abc123xyz"

voices = requests.get(
    "https://audexum.com/api/v1/voices",
    headers={"Authorization": f"Bearer {API_KEY}"},
).json()

for voice in voices:
    print(f"{voice['voice_id']:30s} {voice['language']:10s} {voice['gender']}")

Example output:

text

en_us_female_01                en-US      female
en_us_male_01                  en-US      male
en_gb_female_01                en-GB      female
de_female_01                   de         female
ja_female_01                   ja         female

Pass the voice_id string directly in your synthesis request.

Saving the WAV File

The response body is raw PCM WAV — no base64 decoding needed. Write it to disk with standard file I/O:

python

import requests
from pathlib import Path

def synthesize(text: str, voice_id: str, output_path: str, api_key: str) -> Path:
    response = requests.post(
        "https://audexum.com/api/v1/tts",
        headers={
            "Authorization": f"Bearer {api_key}",
            "Content-Type": "application/json",
        },
        json={"text": text, "voice_id": voice_id},
    )
    response.raise_for_status()
    path = Path(output_path)
    path.write_bytes(response.content)
    return path

# Usage
output = synthesize(
    text="Welcome to our application.",
    voice_id="en_us_female_01",
    output_path="welcome.wav",
    api_key="sk_live_abc123xyz",
)
print(f"Audio saved to {output} ({output.stat().st_size} bytes)")

If you need MP3 instead of WAV, pass "format": "mp3" in the JSON body.

Error Handling

Handle rate limits, invalid voices, and authentication errors explicitly:

python

import requests
from requests.exceptions import RequestException

def safe_synthesize(text: str, voice_id: str, api_key: str) -> bytes | None:
    try:
        response = requests.post(
            "https://audexum.com/api/v1/tts",
            headers={
                "Authorization": f"Bearer {api_key}",
                "Content-Type": "application/json",
            },
            json={"text": text, "voice_id": voice_id},
            timeout=30,
        )

        if response.status_code == 401:
            raise ValueError("Invalid API key. Check your sk_live_ token.")

        if response.status_code == 402:
            raise RuntimeError("Character quota exceeded. Upgrade your plan or wait for reset.")

        if response.status_code == 422:
            raise ValueError(f"Invalid request: {response.json().get('detail')}")

        response.raise_for_status()
        return response.content

    except RequestException as e:
        print(f"Network error: {e}")
        return None

Status Codes Reference

Code	Meaning	Action
200	Success	Write response body to file
401	Bad API key	Check the Authorization header
402	Quota exceeded	Upgrade plan or wait for monthly reset
422	Bad request body	Check voice_id and text fields
429	Rate limited	Retry with exponential backoff
500	Server error	Retry once, then contact support

Batch Processing

For batches of strings, stay under the rate limit with a small delay between requests:

python

import time
import requests
from pathlib import Path

def batch_synthesize(items: list[dict], api_key: str, delay: float = 0.2):
    """
    items: [{"text": "...", "voice_id": "...", "output": "file.wav"}, ...]
    """
    results = []
    for item in items:
        response = requests.post(
            "https://audexum.com/api/v1/tts",
            headers={"Authorization": f"Bearer {api_key}"},
            json={"text": item["text"], "voice_id": item["voice_id"]},
        )
        if response.ok:
            Path(item["output"]).write_bytes(response.content)
            results.append({"file": item["output"], "status": "ok"})
        else:
            results.append({"file": item["output"], "status": response.status_code})
        time.sleep(delay)
    return results

Pricing Comparison

Before committing to a plan, here is how the major TTS APIs compare on price per character:

Provider	Free Tier	Paid Entry	Cost per 1M chars	Card Required for Free
Audexum	10K chars/mo	€4 / 100K	~€8 (PAYG)	No
ElevenLabs	10K chars/mo	~$5 / 30K	~$11	Yes (some plans)
OpenAI TTS	None	Pay-as-you-go	~$15	Yes
Murf.ai	Limited (watermarked)	~$19/mo	Varies	Yes
Google Cloud TTS	1M chars/mo (WaveNet 0)	Pay-as-you-go	$16 (WaveNet)	Yes

Audexum is several times cheaper than ElevenLabs and roughly 2× cheaper than OpenAI TTS at scale. The PAYG option at €8/1M chars is useful for projects with unpredictable volume.

Plan Summary

Plan	Characters/mo	Price	Best For
Free	10,000	€0	Prototypes, testing
Starter	100,000	€4	Small apps, bots
Pro	500,000	€12	Production apps
Scale	2,000,000	€30	High-volume pipelines
PAYG	Unlimited	€8/1M	Bursty or unpredictable load

Next Steps

The full Python implementation fits in under 30 lines. The free tier is enough to build and test a complete integration before you spend anything.

Get your API key at audexum.com/signup and check the API reference for the full list of parameters including speed, pitch, and output format controls.

Related: Cheapest text-to-speech API in 2026
Related: ElevenLabs alternative with a free tier
Related: TTS API for Discord bots

By Petar, founder of Audexum. Running a bootstrapped TTS SaaS gives you strong opinions about API design.