All features

Tell your story

Speaking your story

Dictate, hold a spoken conversation, or upload recordings you already made.

What it is

CharmWriter offers three separate ways to write by speaking — none of them mandatory. Everything on this page is an alternative to typing, never a replacement for it: the chat box accepts text at every step, and you can write a whole book without touching the microphone. Dictation is press-to-talk: you hold a button, speak, and your words appear in the chat box for you to send. Talk to Charm is a hands-free spoken conversation — Charm listens, replies out loud, asks questions, and drafts prose when you say so. Upload a Recording takes audio you already made elsewhere, such as phone voice memos, and turns it into session material. All three feed the same place: your conversation, from which prose is drafted and the Story Bible is built. Charm can also read its own replies back to you, which is the one thing here that needs no microphone at all.

Where to find it: Dictation: microphone button in the chat composer. Talk to Charm: "Talk to Charm" button in the mode bar (enable it first in Settings → Voice). Upload: the upload button beside the microphone, on by default (Settings → Upload a Recording turns it off). Read aloud: the speaker button under any of Charm’s replies. Listen to a chapter: the Listen button at the top of the Chapter Draft panel.

What this covers

  • Press-to-talk dictation into any chat box
  • Live transcript preview while you speak
  • Whisper transcription for accuracy
  • Auto-send after a configurable pause
  • Talk to Charm: hands-free spoken conversation
  • Charm speaks back in your book language
  • Spoken workflow commands
  • Push-to-talk mode for slow thinkers
  • Barge-in — interrupt Charm mid-sentence
  • Voice, speaking speed and pause sensitivity settings
  • Upload existing recordings (up to 90 minutes)
  • Review a transcript before it enters your book
  • Have any of Charm’s replies read aloud — no microphone needed
  • Listen to a whole chapter read aloud, with the cost shown first
  • Your characters read their own replies aloud, each in their own voice
  • Hands-free: talk to a character out loud and they answer out loud

How to use it

  1. 1

    Dictate into the chat box

    Press the microphone in the composer and talk. You see a live preview as you speak, and the accurate transcription lands when you stop. Edit it if you like, then send.

  2. 2

    Turn on Talk to Charm

    Open Settings → Voice and switch on Voice Storytelling. This is a per-device setting, so turning it on on your laptop does not turn it on on your phone.

  3. 3

    Have a spoken conversation

    Tap "Talk to Charm" in the mode bar and grant microphone access. Speak naturally; when you pause, Charm replies out loud with a short question or observation, then listens again.

  4. 4

    Say "write it up" when a scene is done

    That spoken command turns the conversation so far into prose and appends it to your chapter draft. You watch it arrive on the page while it is written.

  5. 5

    Switch to push-to-talk if you think in pauses

    In the voice controls, switch the turn mode from Auto to Push. Charm then waits for you to press Send rather than deciding your pause meant you had finished.

  6. 6

    Upload recordings you already made

    Use the upload button beside the microphone. Pick one or many files, check the quoted cost, and review the transcript before accepting it into your session. If you never upload audio, Settings → Upload a Recording removes the button.

  7. 7

    Have a reply read back to you

    Press the speaker button under any of Charm’s replies to hear it in your book’s language. Press it again to stop. This needs no microphone permission, so it works even if you have never turned on Talk to Charm.

  8. 8

    Listen to a whole chapter

    Open the Chapter Draft panel and press Listen. Charm shows how long the chapter runs and roughly what it costs before anything is spoken, then reads it while you follow along — pause, resume or stop at any point. It reads what is on screen, including edits you have not saved yet. If a passage cannot be read aloud, the chapter stops at that paragraph and offers to carry on from there — it never skips ahead over words you have not heard.

What it gives you

Speech beats the blank page

Most people speak far more fluently than they write. Talking bypasses the inner critic that stalls a first draft, and CharmWriter is built to catch what comes out.

Hands-free means genuinely hands-free

In Talk to Charm you never touch the keyboard. Starting a chapter, finishing a session, asking for a read-back and writing up a scene are all spoken commands.

You can interrupt

If Charm is talking and you want to keep going, just talk. It stops and listens, the way a person would.

Old recordings are not wasted

Voice memos from a car journey or an interview with a relative can go straight in. You review the transcript before a single word joins your book.

It handles being interrupted by life

If a long upload is cut off by a dropped connection or a backgrounded phone, the transcript is saved anyway and offered back to you next time. You never pay twice for the same recording.

Spoken commands in Talk to Charm

While in a spoken session you can say: write it up (draft the conversation into prose), start a new chapter, finish this chapter, finish the session, analyse (run the post-session pipeline), read that back, pause, end session, and help.

The state-changing ones ask for a spoken "yes" before acting, so a sentence of narration that happens to contain the words cannot silently restructure your book. You can also say "go to chapter seven", "make a note: …", and "just between us, …" for an off-the-record question that does not enter the manuscript.

Choosing a voice and a speed

Settings → Voice lets you pick which voice Charm speaks in, how fast it talks, and how long a pause should be before it decides your turn is over. If you tend to think mid-sentence, lengthen the pause sensitivity or switch to push-to-talk.

The speaking speed is applied to the audio itself rather than to playback, which is why it stays clear at higher speeds instead of clipping words.

Good to know

  • Talk to Charm is off by default and has its own switch in Settings. Upload a Recording is on by default and can be switched off there too.
  • Talk to Charm is a per-device setting, stored on the device rather than your account.
  • Reading a reply aloud is billed like any other spoken audio, so a long passage costs credits. In the chat, nothing is spoken until you press the button. In a character window that changes if you ask it to: Speak replies reads every new reply aloud, and hands-free speaks them all by design.
  • Listening to a whole chapter costs roughly half a credit to two credits depending on its length and your book’s language, and takes about as long as reading it. The estimate is shown before anything is spoken, and stopping early means you only pay for what was read.
  • Uploaded recordings are capped at 90 minutes and 100MB per file. Longer recordings should be split before uploading.
  • Transcription is billed by the second of audio, and long uploads are quoted before they run.
  • Speech recognition mishears invented names — a fantasy character no dictionary knows will sometimes come out wrong. That is why uploads get a review step and dictation is editable before sending.
  • Spoken workflow commands are detected from your words only. Charm cannot trigger them, and it never acts on a command it merely mentions.

Questions

Do I have to speak? Can I just type?

You can type everywhere. Every surface that accepts speech also accepts typing, and most writers switch between the two depending on mood and surroundings.

Why does Charm cut me off while I am still thinking?

Auto turn mode ends your turn after a pause. Either lengthen the pause sensitivity in Settings → Voice, or switch the turn mode to Push so it waits for you to press Send.

Can I dictate in Ukrainian or Czech?

Yes. Charm's spoken replies and the drafted prose follow your book language. Speech recognition follows the interface language, so set the interface to the language you speak — which is what lets you talk in one language and have the book written in another.

What happens to my recordings?

Audio is transcribed and the text becomes part of your session. Your recordings and transcripts are yours, are visible only to you, and are not used to train models.

Does the transcript go straight into my book?

No. An uploaded transcript lands in the review panel first, where you can edit it, and it enters the conversation — not the manuscript. Prose is only written when you ask for it.

Start your first chapter

Every new account starts with two free credits, enough to talk a whole chapter through, have it drafted, and see how it reads.

Get started free