Skip to main content
Personalized video — without the recording session

Bonjoro asks you to film 30 videos a day. OCCS asks for your Brand Character.

Train a Brand Character LoRA once, drop a {personalized_video} merge token into any email broadcast, and every recipient gets a unique AI talking-head clip with their first name. Same outcome, no recording booth.

OCCS

Pay per render only. Name-level dedupe means 1000 recipients with 200 unique first names cost 200 renders, not 1000.

Bonjoro

$35-99/mo per host. Solo plan caps at ~30 personal video sends/day; team plans require everyone to record their own.

Source: bonjoro.com/pricing (checked 2026-05-07; per-host send caps + recording requirement)

Cost to ship 1000 personalized videos in a single broadcast

Bonjoro Solo (5h recording)
OCCS (Replicate render × dedupe)
$99
$2

100 unique names

$99
$6

300 unique names

$99
$20

1000 unique names

Bonjoro is human-recorded — the cost isn't subscription, it's the 5+ hours per 1000 recipients spent recording. OCCS bills the Replicate render time only (~$0.02/clip × distinct names). At 100 unique first names you ship the same campaign for $2 of compute and ~15 minutes of pre-render.

Feature-by-feature

Honest comparison. Where Bonjoro ships the same capability we say so — the moat is in the rows where they don't.

Feature OCCS Bonjoro
Per-recipient personalized video
First name spoken in the opener
Embeds in email broadcast
Animated GIF inbox preview (Gmail / Outlook native)
AI-generated face (no recording session required)
Trained Brand Character (LoRA fine-tune, ~$5)
ElevenLabs voice clone (Instant) or preset voice
Replicate latentsync lip-sync
(script × voice × model) cache dedupe
Pricing scales with distinct names, not recipients
Same pipeline reuses for ad-creative talking-head openers
Send-time merge token in any broadcast
Multi-channel platform (email + social + paid + blog)

Bottom line

Bonjoro converts because the video is real. The constraint is human time. OCCS gives you the same per-recipient personalization without the constraint — so you can run it at a 1000-broadcast scale where Bonjoro can't.

Common questions

Won't an AI talking-head feel less personal than a real recording?
It's a different trade. A trained Brand Character is recognizably YOUR character — same face every email, every broadcast, every ad. Recipients learn it the same way they learn a host or a tagline. For high-touch sales (5-10 deals a day), Bonjoro's real human still beats us. For broadcast volume (100s+ a day), the math flips — most teams stop sending personalized video at scale because the recording load is impossible.
How do I train a Brand Character?
Upload 8-20 reference photos of the person you want to clone. We train a Replicate flux-dev LoRA in ~15 minutes for ~$5. After that, the character is reusable across every personalized broadcast, every ad creative talking-head opener, and every Remix Lab scene.
Can I clone my own voice?
Yes. ElevenLabs Instant Voice Cloning takes a 30-second sample. Or use a curated preset if you don't want your voice in production.
What if a recipient's first name isn't in our database?
The merge token expands to an empty string for that recipient — they see the rest of the email body without a video. Same fallback Bonjoro and Mailchimp use for missing merge fields.

Ready to switch?

Free tier, no card required. Run OCCS in parallel with Bonjoro for 14 days — keep what works.