Horror Narrator Voice: AI Voices for Scary Stories
Horror narrator voices heard on real scary stories: Poe's The Raven, The Tell-Tale Heart and an original creepypasta. Then narrate yours, whispers and all.
Last updated: 2026-10-26
A horror narrator voice is low, slow and close to the mic: it holds back, lets the pauses do the work, and drops to a whisper right before the worst line. Below, AI narrators do exactly that on real scary stories, from Edgar Allan Poe's The Raven to an original creepypasta, followed by the tags, effects and settings to narrate your own. Everything on this page is free to hear; narrating your own story uses credits.
Hear the horror narrators
Tap a voice to hear it. To voice your own text, paste it into the workbench: 7 credits per 100 characters, starting with the $3.90 trial pack.
Hear a horror narrator on a short scary story
The shape most horror channels use: a calm setup, one detail that is wrong, and a whisper on the last line. Trailer Voice on Eleven v4, one take:
[whispers] This morning the key was gone. And the door was still locked.Hear it in full horror stories
The theater has eight English horror episodes, seven of them Poe in his own words (a few abridged), every voice generated. Start with The Raven: the scholar narrates, and the Raven itself is Trailer Voice.
The Tell-Tale Heart, told by a killer who keeps insisting he is sane:
Gothic, with British Butler narrating and Trailer Voice as the dead partner dragging his chain up the stairs:
A modern shocker under three minutes long, with Late-Night DJ as the narrator: a sister three years dead calls at 3:17 a.m.
More Poe in the Edgar Allan Poe collection: The Masque of the Red Death, The Cask of Amontillado and The Oval Portrait, plus an original snowed-in cabin story, It Knocks in My Brother's Voice.
Several of those episodes use ElevenLabs default voices (Brian narrates The Raven, Callum The Tell-Tale Heart), which ElevenLabs retires on December 31, 2026 (see ElevenLabs default voices retiring for the replacements). The finished episodes keep playing, but for a new series, start with a voice that stays: the designed voices below.
Horror narrator voices to choose from
| Voice | What it sounds like | Best for |
|---|---|---|
| Trailer Voice | Ultra-deep, gravelly American | Dread, creature features, "something is coming" |
| British Butler | Calm, dry, crisp British RP | Gothic and Victorian ghost stories, Poe |
| Tech Founder | Measured American, quiet conviction | Analog horror broadcasts, found-footage logs, incident reports |
| Ice Queen | Low, slow, every word a blade | Villain monologues, the thing behind the door (designed in Mandarin, so English has an accent) |
| Late-Night DJ | Soft, close-mic Mandarin | Chinese ghost stories such as Painted Skin, rule-based horror |
| Old Storyteller | Gravelly Chinese storyteller | Classic Chinese ghost tales such as Nie Xiaoqian |
None of them is a copy of a real narrator or YouTuber. If none fits, describe the one you hear in your head ("a hoarse old lighthouse keeper who speaks slowly and almost never raises his voice") and design it with prompt to voice for 20 credits. For antagonists, see the villain voices.
How to direct a scary narration
Dread comes from restraint. These are the controls, all written straight into the script:
| Technique | Write it like this | What you hear |
|---|---|---|
| Whisper | [whispers] It was still in the house. | The next words drop to a whisper |
| Fear | [nervous] I told myself it was the wind. | Hesitant, unsteady delivery |
| Breath | [sighs] or [gasps] before a line | A sigh or a sharp intake of breath |
| A long pause | I opened the door... | A held beat before what comes next |
| One beat per line | A new line for each sentence of the reveal | Shorter phrases, more space between them |
The tags work on Eleven v4 and are never spoken or shown in the subtitles. Keep it to two or three per paragraph: a story where every line whispers stops being scary. Four tags in one short scene, a whisper and a nervous question before the joke breaks the tension (Sweet Girl):
[whispers] Okay, don't move. I think it's still in the kitchen. [nervous] What if it heard us? [laughs] Relax, it's just the cat. [sighs] You scared me half to death.A script ready to paste looks like this:
The cabin had one rule: lock the door before dark.
I locked it at six. I checked it at seven.
At eight, I heard the handle turn... slowly.
[whispers] It wasn't trying to get in.
[nervous] It was testing whether I was still awake.
The full tag list, with audio for each, is in the emotion tags cheat sheet.
Analog horror and emergency broadcast voices
Analog horror lives on calm voices through bad equipment: a test broadcast, a phone line breaking up, a machine reading instructions. Add the effect while the line is generated; it costs nothing extra. Tech Founder through the old-radio effect:
The robot and bad-phone effects, for the machine voice and the call that keeps breaking up:
One effect applies to a whole generation, so generate the broadcast lines on their own and cut them into the story. More effects and examples: robot voice generator and phone voice effect. To start with the effect already selected, open the workbench with old radio.
A full cast for horror dialogue
When the story has people talking (the friend on the phone, the voice outside the door), one narrator is not enough. Paste the story into the audio drama studio: it finds every character, gives each a voice, tags the emotion line by line and renders the scene with subtitles. Try it on a short creepy story here (finding the cast needs credits in your account but uses none):
Generating needs credits · trial pack from $3.90 · hear finished dramas in the theater first
Making a horror story video: workflow and cost
- Write for the ear. Short sentences, one beat per line, sounds described rather than shown. For a story with several speakers, follow the audio drama script format.
- Narrate. Up to 5,000 characters per generation in the text-to-speech tab (about six minutes of English), or up to 50,000 in one long-form job with one continuous subtitle track.
- Add the room. Creaking floors, wind, a distant door: generate them with the AI sound effects tool and mix them under the voice.
- Edit. Download the MP3 and SRT and drop both into your editor; the subtitles are already timed.
| Story | Characters | Credits | At list price (new-user first month) |
|---|---|---|---|
| A 1-minute short | ~850 | 60 | $1.25–2.00 ($0.50–0.79) |
| A 10-minute story | ~8,500 | ~600 | $12.50–19.90 ($5.00–7.90) |
| A 30-minute story | ~25,500 | ~1,800 | $37–60 ($15–24) |
Rate: 7 credits per 100 characters for one narrator, the same 70 per 1,000 characters for a full cast. Price ranges are the Creator and Starter monthly plans; a Starter month (600 credits, $7.90 the first month for new users, then $19.90) covers one 10-minute story. The trial pack is 150 credits for $3.90 in the first 24 hours after sign-up (list $9.90). See pricing.
The audio you generate is yours to use commercially, monetised channels included. YouTube's monetization policies exclude "inauthentic" content: AI content "made with generic or unoriginal templates giving the impression of mass production". Narrating stories you wrote, or public-domain classics you adapt and edit, is a different thing from a channel that churns out template videos.
FAQ
What voice do horror narrators on YouTube use? Usually a low, slow male voice recorded close to the mic, with long pauses and a drop to a whisper before the worst line. We do not imitate specific creators: the narrators on this page are voices we designed from a text description, such as Trailer Voice and British Butler.
Can the AI whisper? Yes. Start the line with [whispers] on Eleven v4. The tag is performed, never read aloud, and never shows up in the subtitles. How strongly a voice responds varies, so test the tag on a short line with your narrator first.
How do I get an analog horror or emergency broadcast voice? Generate the line with an effect applied: old radio, bad phone line or robot. The effect is added while the speech is generated, at no extra cost, in the text-to-speech tab of the workbench. A calm, flat delivery sounds more unsettling than a dramatic one.
Can I monetise videos with these voices? Yes. The audio you generate is yours to use commercially under the terms of service. YouTube applies its own rules on top: it does not monetise mass-produced, template-like content, so the story and the edit need to be your own.
How much does it cost to narrate a scary story? Listening is free: every sample and the Poe episodes in the theater play without an account. Narrating uses credits: a 10-minute story is about 8,500 characters, which is about 600 credits, or $12.50–19.90 on monthly plans at list price ($5.00–7.90 in a new account's first month). The one-time trial pack (150 credits, about 2,100 characters) is $3.90 in the first 24 hours after sign-up (list $9.90).