OLED Studio

0%

warming up the voices…

TYPE IT.HEAR IT.SOUND HOME.

Studio-grade Nigerian English from plain text instant for one listener, elastic for one hundred thousand. This is voice you deploy.

Start speaking
0.2s first audio24kHz studio grade∞ pool scale
owambe.py Run
1
2
3
4
5

● Now playing · real render

0:00

Owambe invitation

Adaeze · Nigerian English · 24kHz

“Come chop rice, come dance, come celebrate with us…”

/ The problem

Every synthetic voice sounds like somewhere else. Ours pronounces Owambe correctly. It knows no wahala is a full sentence. It carries the rhythm of home Lagos traffic, Abuja mornings, Port Harcourt nights in every single line.

/ The voices

Three voices.
One homeland.

01 — Warm · Clear

ADAEZE

Announcements, classrooms, customer care. The voice of patience — she slows down for every name.

EDU / IVR / CAREaudition →

02 — Deep · Steady

EMEKA

Storytelling, ads, narration. Unhurried confidence with the weight of evening radio.

ADS / FILM / BOOKSaudition →

03 — Bright · Quick

FUNMI

Creators, skits, social. A smile you can hear — built for timelines that never sleep.

CREATORS / SOCIALaudition →

/ The machine

KEEP
SCROLLING.

The page is pinned. The pipeline moves. Five stops from keystroke to soundwave.

01INPUT

YOU TYPE

A sentence, a script, a saga. Plain text is the entire interface — app, API, or studio.

02GATEWAY

GATEWAY SERVES

One FastAPI gateway validates the key, picks the voice, and renders inline in ~0.5s.

03QWEN

QWEN POOL RENDERS

Pooled Qwen voices render Nigerian English at 24kHz. Scales on Modal, idles to zero.

04WAV

WAV STREAMS

Audio leaves as WAV over HTTP, SSE, or live WebSocket chunks. First audio in ~0.2s.

05OUTPUT

EARS RECEIVE

WAV straight to a player, or into your app over WebSocket. Journey complete.

/ Developers

Three lines
to a voice.

No SDK to learn. HTTP in, audio out then scale the fleet with one environment variable.

import requests

wav = requests.post("http://127.0.0.1:8000/v1/speech",
  json={"text": "Good morning Lagos", "voice_id": "ada"})
open("morning.wav", "wb").write(wav.content) # 24kHz WAV

0s

FIRST AUDIO

0kHz

STUDIO GRADE

0

VOICES LIVE

0 ways

HTTP · WS · SSE

/ Word of mouth

Loved where
it matters.

“It pronounced my grandmother's name correctly on the first try.”
Amara · Storyteller
“Our jingles finally sound like us.”
Tunde · Producer, Lagos
“The Igbo intonation is spot on — my callers stay on the line.”
Jonas · Backend engineer
“It pronounced my grandmother's name correctly on the first try.”
Amara · Storyteller
“Our jingles finally sound like us.”
Tunde · Producer, Lagos
“The Igbo intonation is spot on — my callers stay on the line.”
Jonas · Backend engineer
“Dropped the speech API into our app in an afternoon.”
Priya · Game dev
“First audio lands before I lift my finger off play.”
Ada · Studio user
“My students lean in when Adaeze speaks.”
Mrs. Okafor · Teacher, Abuja
“Dropped the speech API into our app in an afternoon.”
Priya · Game dev
“First audio lands before I lift my finger off play.”
Ada · Studio user
“My students lean in when Adaeze speaks.”
Mrs. Okafor · Teacher, Abuja

/ The journey

Pick a moment.
Press play.

Good morning Lagos. Danfo don dey road, sun don rise, and today go good. This na Adaeze, and I dey here to read your day to you.

128/500