Multi-channel AI outreach is liveSee it in action
On-camera personas

One recording. Every prospect. Your face, your voice.

Record a 90-second take. Outvid trains an on-camera persona that delivers a fresh script to every prospect, in your real footage.

Sends and syncs with the tools you already run

HubSpotSalesforceLinkedInAttioInstantlyWhatsAppClayCalendly

One take, three prospects

What a single recording turns into

The take you record once, and three videos built from it, each delivering a different prospect's script.

Step 1 · the one takeAbout 90 seconds, webcam is fine
The consent recording an on-camera persona is trained from

You record this once. Nothing about your face is generated.

Step 2 · every prospectSame face, new script
  • Marcus · Ramp

    Head of Sales

  • Sofia · Lattice

    VP Sales

  • Raj · Notion

    Head of Growth

Each one is a fresh script, delivered in your trained voice.

Why on-camera, not synthetic

It really is you on the screen

Your real recording stays on screen; only the script and the voice delivery change per prospect.

Visual: your real footage

The video on screen is the recording you uploaded, never a synthetic avatar.

Voice: your trained voice replica

Your voice is trained from the same take, so tone, pacing and accent stay consistent.

Per-prospect script delivery

A freshly written script per prospect, naming their company and their context.
How it works

From a single recording to a campaign-ready persona

A clean webcam recording is enough. No studio, no greenscreen, no agency.

  1. Record one short consent take

    90 seconds on a clean webcam is enough.

  2. Outvid trains and reviews

    Your voice and delivery are trained from that take, then you review the result.

  3. Ship per-prospect videos

    Outvid writes a fresh script per prospect and your persona delivers it.

Recording

A laptop webcam and a quiet room. That is the whole input.

  • Your footagestays on screen
  • Your voicetrained from the same take
  • A fresh scriptwritten per prospect

Your clone does the talking

You don't have to perform. Your clone does.

Your clone delivers the right words for each prospect, in your voice, with no nerves and no stumbles.

Says it perfectly, every time

Every send is a clean, fluent take with no stumbles and no filler.

You never write a word

Outvid researches each prospect and writes the script; your clone delivers it.

Your voice, every language

It speaks in your own voice, in your prospect's language.

It's your clone, you're in control

Review every script and video before it reaches a prospect.

Give your clone a warm take to learn from, and that warmth carries into every send.

One clone, every language

Your clone speaks every language

Record once. Your clone delivers the same message, same face and voice, in your prospect’s language.

Personalized outreach video — preview
Sample script

Daniel — congrats on the funding round.

One take. Switch the language — it’s still you.

+ 17 more languages

Quality & consent first

You approve every persona before it sends

Every recording starts from your consent, every persona is reviewable, and every send is auditable.

  • Consent-first: training only starts from authorized recordings you upload
  • Review-first: every persona is approved before campaigns can use it
  • Per-send auditability: every output is reviewable in the campaign timeline
  • Workspace-scoped: nothing crosses tenant boundaries or gets shared
Consent

Authorized recordings only

Personas only train from recordings you upload. No scraping, no lookalikes.
Review

Approve before send

Approve scripts and rendered output before a single video reaches a prospect.
Fidelity

Performance that holds up

Tuned for natural cadence and facial timing on real footage, not synthetic avatars.
Voice

Your voice, not a stock voice

Trained on your own audio, so tone and accent stay yours on every send.

On-camera vs synthetic avatar

Why prospects skip the AI avatar

A prospect who spots a digital twin stops watching the pitch and starts watching the software.

Synthetic avatars

  • A recognisable AI face that prospects spot in seconds
  • Mannerisms feel off, so attention goes to the tool
  • Brand and personal authenticity get diluted
  • Stock avatars, so your video looks like everyone else's

On-camera personas

  • Real footage of you on screen. No AI face
  • Your voice, your tone, your expressions, at scale
  • One recording personalises infinitely without re-shooting
  • Different from teams sending stock-avatar video
Where your persona lands

The page your prospect opens

Emailed, it opens a page with your persona already playing. In chat it plays right in the thread.

Email

A GIF in the message. Clicking it opens their own page.

LinkedIn

The video plays right in the thread.

WhatsApp

Arrives as an attachment in the chat.

Instagram

Lands in their DMs, ready to play.

What's customized:Founders lead with one sharp reason a prospect should care, backed by the proof that's true for their segment.

Sample prospect data, shown for illustration.

Frequently asked

No. The video on screen is your real recording — Outvid never substitutes a synthetic AI clone or AI avatar in your place. Only the script and voice delivery change per prospect; the rest is the take you uploaded.

Synthetic AI avatars generate a fully digital face from scratch — prospects can usually tell, and reply rates suffer (the uncanny valley). On-camera personas keep your actual recorded video as the visual anchor and only personalize the visual delivery. Same speed and scale, no synthetic face on screen.

Most teams are campaign-ready in under an hour after uploading a clean 90-second consent recording.

Yes. The visual on screen is your real recording. Outvid only personalizes the script and the voice delivery — never your face.

A modern laptop webcam and a quiet room is enough. No greenscreen, no studio, no agency.

Always. Every script is reviewable, every render is approvable, and every send is auditable in the campaign timeline.

Yes. Personas are built only from recordings you explicitly upload. There is no scraping, lookalike training or third-party likeness use.

No. Your AI clone does the talking. Outvid writes each prospect's script and your clone delivers it in your voice, cleanly and fluently, with no stumbles and no nerves, on every send. You never have to perform live, improvise, or record per prospect.

Especially you. The hard part of video outreach is doing it well, over and over, at scale, and that's exactly the part your clone handles. You set it up once, and from there it carries your presence to every prospect for you.

Yes. Your clone is trained on your own voice, so your tone and accent stay yours across every send. Give it a warm sample to learn from and that warmth carries through. It delivers the best-written version of your message, in your voice, every time.

Record once. Send infinite personalized videos.

Record a 90-second take and start sending per-prospect campaigns in your real face and voice.

A

Alex Masson

Promotions

Quick question about your pipeline

9:12 AM

Hi there, I came across your company and wanted to see if you're open to a quick 15-minute call about how we help teams...

AI-generated sales video avatar

Your Name

Video Message

Made this video for you, Marcus

Just now
Prospect company research preview
0:47

Hey Marcus, saw Northwind just opened its second site, congrats!

Click to watch your personalized video →

NEW