From an idea to a narrated episode.

OnePress writes the script, picks the voices, and mixes the episode. You never touch a microphone — you review the way you review everything else here: by reading, listening, and saying what to change.

What you get

The script, in full

The episode is written before it's voiced — intro, segments, back-and-forth, outro. It stays in the conversation, so you can read every word before you commit a single second of audio to it.

Narrated audio, ready to play

Voices are synthesized and mixed into a finished MP3. Play it right in the app, or take the file and put it wherever podcasts go — it's a normal audio file, nothing proprietary.

One voice, or a real conversation

A steady monologue, or a host and a guest trading lines with natural back-and-forth — the script is written for whichever you choose. Want quiet background music under the voices? Say so; it gets mixed in.

How it works

  1. 1

    Describe the episode

    Topic, audience, and length — "a five-minute weekly recap of the EV market for busy executives" is a complete brief. Rough is fine; the agent will sharpen it with you.

  2. 2

    The agent asks the key questions

    Format and tone — one voice or two, interview or essay, serious or playful. A short exchange that decides what the episode sounds like before any audio exists.

  3. 3

    Research, if the episode needs it

    Facts in the script come from the same web research as everything else on OnePress — searched, read, and checkable, not asserted.

  4. 4

    Script first, then sound

    The script gets written and you can read it. Then each part is voiced and the tracks are mixed into one episode — you see it happen, step by step.

  5. 5

    Listen, then refine

    "Punchier intro." "Swap the guest voice." "Trim the middle section." Your notes go in as ordinary chat messages, and the episode rebuilds around them.

Voices

There's a curated voice roster, matched to your episode's language — native Mandarin voices for Chinese episodes, native English voices for English ones, so nobody sounds like a foreigner reading a script. The agent picks a sensible pair by default and tells you which ones it used; swap them with one word.

Or narrate it yourself: record each segment and drop the audio files into your workspace — the agent mixes your real voice over the music bed. Voice cloning is intentionally not offered: a catalog voice you choose beats a synthetic copy of you.

Automate it

A daily news podcast in your niche, without a studio or a schedule you have to keep:

"Every weekday at 7, a four-minute episode on what happened overnight in the robotics industry — two voices, sources from the run."

The agent researches, writes, and voices a fresh episode on its own schedule, and every run lands as a conversation you can read and listen to. See Automations for how schedules work.

Frequently asked questions

Do I need to record anything?

No — every voice in the episode is synthesized from the roster. The one exception is voice cloning: if you want the episode narrated in your voice, you upload one short recording once, and it's reused from then on.

How long can an episode be?

You set the length in the brief — "a five-minute recap" produces a script that runs about five minutes. Longer episodes simply mean longer scripts and more credits to voice them.

What do I end up with?

Two things: the script — readable, and revisable by chat like any OnePress work — and the finished MP3, playable and downloadable. Change the script, and the audio follows.

No mic, no timeline, no editing

Open the app, describe the episode, and listen to what comes back. Change what you don't like; keep what you do.

Open Podcast