Screenwriter Workflow

Dictation for Screenwriters:
Use Voice in Final Draft on Windows

Final Draft's built-in voice dictation is Mac-only. On Windows, you have been stuck. StarWhisper fills the gap, typing dialogue, action, and parentheticals straight into Final Draft, WriterDuet, Fade In, and Celtx as you speak.

Download for Windows
Microsoft Store
  • Trusted by Windows
  • Quick 30-second setup
"She steps into the rain and pulls her hood up, scanning the empty street..."

The Windows screenwriter's missing voice layer

Speaking dialogue out loud is a screenwriting tradition. Dictating it directly into Final Draft makes the tradition practical.

For screenwriters on Windows

Talk the script, format with Final Draft

Speaking dialogue out loud catches what reads naturally and what does not. With dictation, the "say it aloud" pass IS the writing pass. Drafts move faster.

  • Works in Final Draft, WriterDuet, Fade In, Celtx
  • Dialogue, action, parentheticals, all fields
  • Local Whisper, scripts never leave your machine
  • Faster than Dragon, cleaner than Windows Voice Typing
  • Free 500 words/day, $10/month for unlimited
What dictation does not do

It is not a screenplay formatter

StarWhisper types what you say. The element switching, the slug line caps, the character cue formatting, all that is Final Draft's job. Dictation handles the words; the script app handles the format.

  • Slug lines: still type "INT." and tab
  • Character cues: type the name, tab to dialogue
  • Margin formatting: Final Draft handles it
  • Dialogue and action blocks: dictate
  • Long parentheticals: dictate

Where dictation fits in a screenwriter's day

Six surfaces in a working script where voice replaces typing

Dialogue lines

Dialogue is the surface where dictation shines. You are already supposed to say lines aloud to test how they read. Dictating bypasses the typing tax and lets the spoken rhythm flow straight onto the page, exactly the way screenwriters have always taught.

Action and scene description

The action block is the densest prose in a screenplay. Twenty to forty words per beat, often a dozen beats per scene. Dictating this at 130 to 150 words per minute beats typing at 60. Cumulative time savings across a draft are measured in days, not hours.

Parentheticals and stage directions

Long parentheticals like (sotto voce, eyes drifting toward the window) and inline stage directions add texture but are slow to type. Dictating them keeps your hands on whatever you are already doing and your eyes on the screen, not the keyboard.

Treatments and outlines

Before the script comes the treatment, the beat sheet, the synopsis, the logline. These live in Word, Scrivener, Notion, or Google Docs depending on your workflow. The same dictation hotkey works in every one, so you stay in one input mode end to end.

Production paperwork and notes

Once the script is out, you are writing pitch documents, character bios, episode breakdowns, and notes back to producers. None of that is screenplay-formatted, all of it is long prose. Dictation cuts the typing time roughly in half.

Email to your agent, manager, and writers' room

The screenwriter's inbox is its own job. Replies to notes, scheduling for table reads, pitch follow-ups. Dictation works the same in Outlook, Gmail in a browser, and Slack as it does in Final Draft, because the layer is below the application.

Why Windows screenwriters have been stuck without good voice tools

Final Draft is the industry-standard screenwriting application. It is also one of the longest-running ones, and its feature set reflects decades of mostly-Mac-centric development. The built-in voice dictation features tie into Apple's system dictation framework, which is why they have always worked on the Mac version and never had a real Windows equivalent. Some marketing copy across the years has implied parity, but in practice Windows screenwriters who want voice input have had a poor set of options.

Dragon NaturallySpeaking was the historical default. It works, but it is expensive at $699 for a one-time license, the user interface has barely changed in years, and active development has slowed dramatically since Nuance was acquired by Microsoft. The built-in Windows Voice Typing is free and tolerable for chat-style sentences, but its accuracy on long creative prose is uneven, and it does not handle the rhythm or vocabulary of dialogue particularly well.

The third option is a modern Whisper-based desktop layer. StarWhisper runs OpenAI's Whisper locally on your Windows machine and types the transcription into whatever text field is focused. Final Draft, WriterDuet, Fade In, Celtx, your beat sheet in Notion, your email to your manager. The same hotkey works everywhere because the dictation happens at the operating system input level, not as a plugin.

The screenwriter's classic technique, automated

Every screenwriting craft book that talks about dialogue makes the same recommendation: read your scenes aloud. The page is silent; the screen is not. Dialogue that reads on paper as elegant and natural often turns out, when spoken, to be clunky, awkward, or rhythmically wrong. The "say it aloud" pass is a quality gate that catches problems no silent read-through will catch.

Voice dictation collapses that pass into the writing pass. When you say a line of dialogue out loud as you write it, you hear it. If it is clunky, you stop, rephrase, dictate the better version. The bad first draft never makes it to the page in the first place. Many screenwriters who try this report that their dialogue gets better faster, not just because dictating is fast but because the silent typing layer never caught what their ear catches in real time.

Action lines benefit from the same effect for a different reason: pacing. A line of action that reads ploddingly when typed often reads brisk and clear when dictated. You hear the rhythm in your own voice before it lands on the page, and you naturally shape it tighter.

How dictation slots into a Final Draft session

The workflow is undramatic, which is the point. Open Final Draft. Open your script. Click into whatever element you are working on: scene heading, action, character, dialogue, parenthetical, transition. Hold the StarWhisper hotkey, speak, release. Final Draft sees the text as if you had typed it. The auto-formatting that turns "INT. APARTMENT, NIGHT" into a proper slug line still happens; the auto-cap on character names still happens; the margin and indent rules still apply.

For dialogue specifically, the most natural pattern is: type or click into the character cue field, type the character name (it auto-caps), press Tab to drop into the dialogue line, hold the hotkey, speak the line, release. The dictated text lands in the dialogue element with the right margins. Press Tab or Enter to end the dialogue and move to the next element, exactly as you would with typed input.

For action blocks, the pattern is even simpler: click into the action element, hold the hotkey, speak the description, release. Speak the next beat and dictate again. Most screenwriters find that they want to dictate one beat at a time rather than a whole paragraph at once, because that gives them a chance to hear what just landed before committing to the next beat.

Beyond Final Draft: WriterDuet, Fade In, Celtx, and the rest

Final Draft has the most market share in feature television and studio film work, but it is not the only screenplay app on Windows. WriterDuet has a desktop client and a strong web app, and its collaboration features have made it popular for writers' rooms working remotely. Fade In is a clean, fast, well-priced alternative that screenwriters who value craft over enterprise features often prefer. Celtx is widely used for indie and student work and runs in a browser as well as a desktop app. KIT Scenarist and Trelby exist for niche workflows.

StarWhisper works in all of them on Windows without any per-app configuration. The dictation layer types into whatever field is focused. WriterDuet's web app gets dictation into the script editor the same way as Final Draft's native app. Fade In's element switching is different from Final Draft's, but the typed input is the same, so dictation feels identical. Celtx in the browser is just a web text field. Highland 2 is Mac-only as a desktop app and therefore out of scope for this page, but its web tools accept dictation in a browser like any other site.

For the prose drafts that live outside the script, like treatments in Microsoft Word or beat sheets in Scrivener, the same hotkey works. See the voice to text for writers overview and the voice to text for copywriters overview for adjacent workflows that share the same tool.

Whisper accuracy on script-specific vocabulary

The biggest accuracy concern for screenplay dictation is not industry jargon but character names. A modern thriller might have a "Detective Kovalenko" and a "Sister Adelaide" and a "Mr. Renworth," none of which Whisper has any reason to know in advance. In practice, common Western names land cleanly: Sarah, James, Maria, Daniel, Olivia, Marcus, and so on come out correctly. Less common names sometimes need a one-word correction the first few times they appear, after which you learn to pronounce them the way Whisper transcribes cleanly.

Industry jargon (cut to, dissolve to, fade in, fade out, intercut, montage, V.O., O.S., POV, beat, smash cut) lands correctly with Whisper because these words are well represented in the broader English corpus the model was trained on. Slang, regional dialect, period-specific vocabulary, and invented sci-fi words follow the same pattern as character names: a few corrections at first, then it settles down.

Pro users on NVIDIA GPU paths run the medium or large Whisper model, which handles edge cases noticeably better than smaller models. For a dense scene with a lot of unfamiliar proper nouns, the GPU path is worth having.

Privacy: your unproduced script should stay on your machine

An unproduced screenplay is a sensitive document. It has commercial value in a way most prose drafts do not, and the writer often does not want it sitting on a third-party server while in development. The standard cloud transcription model, where audio is uploaded to a vendor, processed remotely, and deleted on some schedule, is awkward for working screenwriters.

StarWhisper runs Whisper locally on your CPU or GPU. The audio never leaves your machine. There is no transcription cloud, no audio retention period, no third-party vendor with access to your dialogue. If you take your laptop to a coffee shop with no WiFi, dictation still works because nothing is uploaded.

Cloud Mode, which sends audio to the OpenAI Whisper API for faster turnaround on weaker hardware, is opt-in and disabled by default. For a working screenplay, leave it off. On any modern NVIDIA GPU the local path is fast enough that there is no performance reason to enable Cloud Mode.

Setup for the first day of voice-driven screenwriting

The setup takes about three minutes. Install StarWhisper from the download page or the Microsoft Store. The installer auto-detects whether you have an NVIDIA GPU and picks the matching pack: CPU, CUDA 11, or CUDA 12. First launch downloads the Whisper model files, which takes a couple of minutes on a normal connection. After that the app lives in your system tray and listens for a hotkey.

For screenwriting, pick a hotkey that does not collide with anything in Final Draft's keyboard map. Right-side modifier keys (Right Ctrl, Right Alt) are good defaults because Final Draft's own shortcuts use the left-side modifiers. If you already use a foot pedal or programmable mouse for video editing or transcription work, those buttons are excellent hotkey targets because they keep your hands on the keyboard for the formatting work that still wants typing.

Test by opening a blank Final Draft document and dictating a few lines of dialogue and a paragraph of action. Once that flow feels comfortable, switch to your working script. For the first week, treat dictation as your dialogue tool only; that is where the win is largest and the learning curve smallest. Once dialogue dictation feels natural, extend to action lines, then to treatments and outlines, then to the email and Slack work that fills the rest of the writer's day. For complementary tools that work everywhere on Windows, see the works everywhere feature overview.

Frequently Asked Questions

Does Final Draft have built-in voice dictation on Windows?
Not in any meaningful way. Final Draft's built-in voice features have long been macOS-only, relying on Apple's system dictation framework. The Windows version of Final Draft does not ship an equivalent voice input feature. Windows screenwriters who want to dictate dialogue and action have historically reached for either Dragon NaturallySpeaking, which is expensive and stagnant, or the built-in Windows Voice Typing, which is mediocre on creative prose. A modern local Whisper desktop layer like StarWhisper fills the gap by typing directly into Final Draft as you speak.
What about WriterDuet, Fade In, and other screenwriting apps?
All of them work with StarWhisper because the dictation happens at the Windows input layer, not as a plugin. WriterDuet's desktop and web app, Fade In on Windows, Celtx, KIT Scenarist, and Trelby all accept dictated text into any focused text field. That includes the dialogue line, action description, parenthetical, scene heading, and character cue. Highland 2 is Mac-only as an app, so it does not apply, but its web companion accepts dictation in a browser the same way any web app does.
Can I dictate scene headings like INT. APARTMENT - NIGHT?
You can, but it usually takes a small workflow adjustment. Whisper transcribes spoken English, so dictating 'INT. APARTMENT, NIGHT' tends to come out as 'Interior apartment night' or similar. Most screenwriters dictate scene headings as plain English and let Final Draft's auto-formatting handle the slug line shape. Alternatively, type the scene heading by hand (it is usually only a few words) and dictate the much longer action and dialogue that follow. The dictation win is in the long blocks, not the slug lines.
What about character names in CAPS above dialogue?
Final Draft and most other screenwriting apps auto-format character cues to caps when you tab into the character field and type the name. The cleanest pattern is to type the character cue by hand, since it is a short word and the auto-format requires being in the right element type. Then tab to the dialogue field and dictate the dialogue itself, which is where the time goes. Some writers dictate the full block including the character name and let Final Draft's smart parsing handle it, but that is hit or miss.
Does it punctuate properly for screenplay format?
Whisper outputs standard English punctuation when you speak it: 'period', 'comma', 'exclamation point', 'question mark', 'new paragraph', 'new line', and so on. For dialogue, that is exactly what you want. Action lines work the same way. Parentheticals are usually short enough that you can type them by hand. The format-specific punctuation, like the dot after INT. or the dash in a slug line, is handled by Final Draft's auto-formatting rather than by your spoken input.
Can I dictate stage directions and action lines?
Yes, and this is where dictation pays off most. Action lines are dense, descriptive prose that takes a long time to type but flows naturally when spoken. 'She steps into the rain and pulls her hood up, scanning the empty street for the man who said he would meet her here.' That is 25 words; typing it takes around 25 seconds; dictating takes around 10. Multiply across a 110-page screenplay and the cumulative time savings on action lines alone are significant.
Does it work with other writing software too?
Yes. Because StarWhisper sits at the Windows input layer, it types into any focused text field in any application. That includes Microsoft Word and Scrivener for prose drafts, Notion and Obsidian for outline and beat sheet work, Google Docs for collaborative drafts, Final Draft and the other screenwriting apps for the formatted script, and email or Slack for production correspondence. Many screenwriters move between several of these tools in a single day, and the same dictation hotkey works in all of them.
Is there a free trial?
Yes. StarWhisper has a free tier of 500 words per day and 3,500 words per week, which is enough to dictate a couple of pages of dialogue or action per day. The Pro plan is $10 per month or $80 per year for unlimited dictation, with a 7-day free trial of the full Pro experience. For a working screenwriter who routinely drafts five or more pages a day, the Pro plan is what you want. For occasional use or learning the workflow, the free tier covers it.

Dictate Your Next Script on Windows

Free 500 words per day. Works in Final Draft, WriterDuet, Fade In, Celtx, and any text field.

Download StarWhisper for Windows