Create and name speakers
Add as many speakers as your video has. Each gets a unique name, color, and avatar used across the timeline, editor, and preview — so characters are always easy to distinguish.
Multi-speaker dubbing • AI voice library
AI Voice Selection — Multi-Speaker Dubbing & Voice Style
Name every voice in your video, browse a large library of natural AI voices, set age character and a custom style prompt per speaker, and fine-tune expression line by line — all in one dubbing editor.
Start for free
🎭 50+ voices per language
🌍 50+ languages
One workspace for speakers, voices, and performance
Manage every speaker in one panel — assign AI voices, set age character, add an avatar, and control tone from a speaker-wide default style down to individual lines. Built for multi-character videos and editorial teams who need precise per-speaker voice control.
👥Speaker management
Name, color, and manage every speaker
Identify each voice in your video as a named speaker. Color-code them on the timeline, track their segment count and speaking time, and assign an AI voice to each one independently.
Segment stats at a glance
Each speaker card shows segment count and total speaking time, so you can spot imbalances and prioritize voice assignments quickly without counting manually.
Search and filter speakers
Find any speaker instantly in large multi-character projects with a search bar that filters the speaker list by name in real time.
Safe delete with segment reassignment
Removing a speaker who has lines prompts you to reassign those segments to another speaker first — no lines are lost or left with a missing voice.
Browse the AI voice library
Filter the voice catalog by gender, personality tags, and your project's target language. Listen to a sample before assigning — so you hear exactly how the voice will sound in your dub before committing.
Gender and personality filters
Narrow the catalog by male or female voice, then filter by personality — Warm, Energetic, Professional, Calm, Mature, Cheerful, Expressive, and more — to find the right fit for each character.
Language-matched voices
The catalog highlights voices that support your project's target language, so you always know which ones will work for your dub without extra filtering steps.
Preview before you commit
Play a sample directly in the voice picker. Age-specific samples play when they exist — so previewing a Young or Teenager voice actually reflects that age character.
🎭Age character
Match every voice to the character's age
Set an age character per speaker to get age-appropriate voice previews and AI dubbing output. Useful for animation, educational content, or any video with characters of different ages.
Adult
The default character — mature voice with full range of personality styles. Covers most narration, interview, and dialogue use cases.
Young
A younger-sounding variant of the assigned voice. Previews and generated audio reflect a young adult or older teen character where a variant is available.
Teenager
Voices the speaker as a teenage character. Age-specific samples play in the catalog so you hear the difference before assigning.
Kid
For child characters in animation, educational content, or children's media. Keeps voice identity consistent with the chosen catalog voice.
Assign an avatar to each speaker
Put a face to every voice. Crop an avatar directly from your video or upload a photo — it appears on speaker cards, the subtitle timeline, and the inline speaker dropdown while editing, so multi-character projects stay readable at a glance.
Crop from video
Scrub to a clear face frame, drag a circle over the face, and resize to fit. The editor opens at the speaker's first segment so you land near the right frame immediately.
Upload a photo
For audio-only projects or when a clean video frame is not available, upload a JPEG, PNG, or WebP photo and crop it with the same circle tool.
Smart fallback display
No custom avatar set? The speaker automatically shows their assigned voice's catalog image, or a colored initial letter — so the timeline is always readable without manual setup.
Voice style — presets & custom prompts
Define a default performance for each speaker. Choose a built-in preset or write your own prompt describing tone, pace, accent, and emotion. The style applies to all of that speaker's lines unless overridden on a specific segment.
- Built-in presets: Calm, Warm, Professional, Enthusiastic, Dramatic, Patient Teacher, Sincere, Urgent, and many more
- Custom presets: save your own named style prompts and reuse them across future projects
- Apply scope: apply to all segments, or only those without a custom per-line style already set
- Style indicator: active preset name appears on the speaker card so you always know what style is active
🎨Per-line control
Fine-tune every line with expression & effects
Override the speaker's default style on any individual line. Or apply the same expression to every line a speaker delivers — with one click.
Expression per segment
Set emotional delivery on any line: Happy, Whispering, Curious, Emotional, Stressed, Sarcasm, Shouting, Scared, Calm, Angry, Excited, Sad, and more.
Acoustic voice effects
Apply scene acoustics per line: Over Phone, Echo Hall, Walkie Talkie, Underwater, Robot, Lo-Fi, Megaphone, AM Radio, Cave, Intercom, and more.
Bulk apply per speaker
Set an expression or effect on every line for one speaker in a single step — other speakers are not affected. Ideal for consistent characters across long projects.
Stacks with voice style
Per-line expression stacks on top of the speaker's Voice Style — so a narrator set to Professional can still have individual Excited or Whispering lines.
Smart workflows that keep you moving
The platform guides you at each step — from assigning voices before generation to previewing exactly what you are getting before committing.
Voice required before generation
If you generate voiceover for a speaker without an assigned voice, the platform prompts you to set one first — so nothing generates with a missing or placeholder voice.
Auto-assign new speaker to current line
When you create a new speaker while editing a subtitle, the editor asks whether to assign them to the line you were on — so new characters are wired up without extra steps.
Three-tier voice preview
Hear a catalog sample before assigning a voice, the original video audio from a speaker's first segment before generating, or the actual AI-generated output after generation — all from the speaker card.
🎙️
Multi-speaker dubbing
Cast Every Voice in Your Video
Assign professional AI voices, set character styles, and fine-tune delivery line by line — natural AI voices for multi-speaker dubbing in one editor.
Natural AI voices • 50+ languages • Studio-quality AI dubbing
AI Voices & Speaker Management — Frequently Asked Questions
How do I assign different AI voices to multiple speakers in a video?
Open the Speakers panel in your project, add a speaker for each person in the video, then browse the AI voice library to assign a unique voice to each one. You can filter by gender, personality, and your project's target language, and play a sample before confirming. Speakers are color-coded on the timeline so you can identify them at a glance.
What is the difference between Voice Style and per-line Expression in AI dubbing?
Voice Style is a speaker-level default — a preset or custom prompt that shapes how the AI performs all of that speaker's lines (tone, pace, emotion, accent). Per-line Expression is an override on a single segment — for example Whispering, Shouting, or Excited — when one line needs a different delivery. Both stack: a narrator set to Professional can still have individual Excited or Scared lines.
Can I preview an AI voice before generating the full dub?
Yes. Every voice in the catalog has a playable sample in the voice picker — listen before assigning. The age setting affects which sample plays, so a Young or Teenager voice previews its age variant. After generation, you can also hear the actual AI output for any line without leaving the editor.
How do age settings (Adult, Young, Teenager, Kid) affect AI voice output?
The age setting controls which voice character variant is used in previews and voice generation. Choosing Young, Teenager, or Kid selects a younger-sounding version of the assigned voice where available — keeping the character consistent with the person on screen without switching to a completely different voice. The age badge appears on speaker cards so the setting is always visible.
Can I use a custom voice style prompt for each speaker?
Yes. The Voice Style field accepts a free-text prompt — for example "Calm and measured, with a slight British accent" or "Enthusiastic and fast-paced". You can also pick from built-in presets or save your own named presets to reuse across projects. When applying, you choose whether to apply the style to all segments or only those without an existing custom per-line style.
Do AI voices sound natural for multi-speaker dubbing?
Yes. The platform's AI voices are designed for natural, human-like speech with emotional nuance, appropriate pacing, and consistent character across all lines. Voice Style, Expression, and age settings give you fine-grained control — so a narrator can sound authoritative while a secondary character sounds casual and energetic, all within the same project.
Start for free → · Video dubbing workflow · Subtitle editor · Enterprise