Takoตะโก้

v1 + v2

A narrator with the warmth of someone telling you the story in person.

Female
Thai, English, and Mandarin Chinese
Friendly reader
takopaxa-tts-flash-v2

ยามเช้าในกรุงเทพฯ แสงแดดอ่อน ๆ ทาบลงบนถนนที่เริ่มคึกคัก

Cast it for#

Tako sits between a narrator and a friend explaining something. The register is light and approachable without turning juvenile, which is the range most explainer and recap work lives in.

It carries long material better than the bright product voices and stays lighter than the documentary reads. That middle ground is where fiction audiobooks and channel narration usually want to be.

Fiction audiobooks: Novels and short fiction, where a narrator holds a listener for hours without adopting a persona.
Explainers and recaps: Channel narration that has to sound like a person and still cover ground quickly.
Guided walkthroughs: Product tours and how-to material where a formal read would feel oversized.

Reach for instead

Khanom Krok
Documentary and heritage material wants an older register that carries the weight for you.
Foi Thong
The same long-read stamina in an even, more neutral register.

In production#

Audiobooks

Fiction and narrative non-fiction.

E-learning

Lesson narration that stays approachable.

Creators

Channel voiceover and video essays.

Call it from the API#

The voice field takes the id above. Every other value in this request is a Paxa TTS Flash v2 default.

Model: Paxa TTS Flash v2
Per request: 5,000 characters
Rate: 10 credits per 1,000 characters
Free credits: 100 credits, about 16 minutes

API voice reference

curl -X POST https://api.paxalabs.com/v1/tts \
  -H "Authorization: Bearer $PAXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "ยามเช้าในกรุงเทพฯ แสงแดดอ่อน ๆ ทาบลงบนถนนที่เริ่มคึกคัก",
    "voice": "tako",
    "model": "paxa-tts-flash-v2"
  }' \
  --output tako.mp3

import { writeFile } from "node:fs/promises";

const response = await fetch("https://api.paxalabs.com/v1/tts", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.PAXA_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    text: "ยามเช้าในกรุงเทพฯ แสงแดดอ่อน ๆ ทาบลงบนถนนที่เริ่มคึกคัก",
    voice: "tako",
    model: "paxa-tts-flash-v2",
  }),
});
if (!response.ok) throw new Error(`TTS failed: ${response.status}`);
await writeFile("tako.mp3", Buffer.from(await response.arrayBuffer()));

import os
import requests

response = requests.post(
    "https://api.paxalabs.com/v1/tts",
    headers={"Authorization": f"Bearer {os.environ['PAXA_API_KEY']}"},
    json={
        "text": "ยามเช้าในกรุงเทพฯ แสงแดดอ่อน ๆ ทาบลงบนถนนที่เริ่มคึกคัก",
        "voice": "tako",
        "model": "paxa-tts-flash-v2",
    },
)
response.raise_for_status()
with open("tako.mp3", "wb") as f:
    f.write(response.content)

Questions#

Does it work for fiction?

Yes. The register keeps a narrator's distance while still sounding like a person, which is the balance long fiction asks for.

How long can one request be?

A single request takes up to 5,000 characters. A book is sent chapter by chapter, or streamed through the live socket while the text is still arriving.

How old does it sound?

It reads as a narrator in their twenties or thirties. When the part calls for an actual character voice, cast Luk Chup or Bua Loi.

Hear it read your own words. New accounts start with 100 free credits. Open the playground and this voice is already selected.

Start free