Full Time
$400/month + monthly performance bonus
40
Oct 2, 2026
EXPERIENCED ONLY. This is not a training position.
I am hiring someone who is ALREADY making AI talking-avatar UGC videos with OMNI inside Google Flow, for paying clients, right now. If you have never used Omni, please do not apply. If you have only used Veo in Flow, Kling, HeyGen or Higgsfield, please do not apply. If you "know CapCut" but have not made AI avatar videos, please do not apply. I will not reply.
I've attached an example video. That is exactly the style and quality I need. Watch it before you apply.
MINIMUM EXPERIENCE (all of these, no exceptions)
1. Hands-on experience generating talking clips with OMNI in Google Flow. This is the main tool of the job. Not Veo only, not another app.
2. At least 3 months of paid work making AI UGC / AI avatar videos for a client or brand.
3. At least 50 finished AI videos delivered. Not clips, finished videos with captions and music.
4. Real CapCut skill: captions, keyframes, audio ducking, speed ramps, without me explaining them.
5. You have already kept ONE character consistent (same face, outfit, voice) across multiple videos. Show me.
WHAT YOU MUST ALREADY KNOW IN OMNI
If any of this is new to you, this job is not for you yet.
- Using a start frame in Frames mode (not Ingredients mode) so the clip starts on the exact image
- Writing short dialogue prompts: the movement, the speaker, the delivery and the exact line, nothing extra
- Picking the right clip length (4s / 6s / 8s) from how long the line is, so words never get cut off and there is no dead air
- Getting the avatar to say the line word for word, with no doubled or skipped words
- Rewording a clip that gets blocked by the filter ("reputational risk") while keeping the dialogue exactly the same
- Generating x1, one clip at a time, and checking each one before the next
THE WORKLOAD
7 finished product videos per day, each 1 to 2 minutes long. I send the script, the reference video, the character reference image and the start frames the day before. Week 1 is 5 to 6 a day while I audit your workflow. After that it's 7.
If you can only do 3 to 4 a day at good quality, say so honestly. I would rather know now than in week two.
WHAT YOU'LL BE DOING
- Generating talking AI avatar clips with Omni in Google Flow
- Assembling in CapCut: cuts, captions, timing, subtle background music
- Recreating a reference video 1 to 1 while swapping in my avatar, pet, product and setting
- Checking every clip before it goes in
- Exporting vertical 9:16, 1080x1920 MP4 and uploading each one as you finish
THE 8 THINGS AN EXPERIENCED EDITOR ALREADY DOES
1. Character consistency. Same face, outfit, pet and setting across every clip and every video.
2. Voice consistency. One voice, no echo, no reverb, no accent appearing mid-video.
3. Lip sync. Off by half a second means regenerate.
4. Captions. Always on, clean, synced, styled like the reference.
5. Background music. Subtle, under the voice, matching the reference.
6. Match the reference. Framing, pacing, energy, shot order. Phone footage look, not cinematic. My setting, not the inspo's.
7. Export quality. 720p from Flow, my CapCut settings, 9:16 1080x1920, no blur, no watermarks.
8. Same-day revisions. Keep your Flow generations and CapCut project so a small fix stays small.
FLOW CREDIT DISCIPLINE
- Images cost 0 credits. Perfect the start frame first, then generate video.
- Omni clips cost credits by length. Never use 10s. Pick the shortest length that fits the line.
- x1 only. Never x2, x3 or x4. Never 16:9.
- Bad clips cost the same as good ones. Get the prompt right before you hit generate.
- Screenshot the credit balance at the start and end of every assignment.
Accounts and credits are provided. Burning them carelessly is the fastest way off the team.
HOW I VERIFY EXPERIENCE (every applicant who gets through)
- You show me your Omni generations in your Flow history on a screen share or a short screen recording.
- You make ONE paid test video with Omni from my script and assets before you are hired. Same quality bar as the real job.
- I get the same recycled portfolio folders from different people every week. If you can't prove the work is yours, you're out.
AUTOMATIC REJECTION
- No Omni experience
- No links to AI videos you personally made
- Only regular editing, motion graphics or stock-footage work
- Portfolio site or Google Drive of work you can't prove is yours
- Does not start with the word OMNI
- AI-written copy-paste application that doesn't answer the questions below
CONSISTENCY
Same hours, same days, every week, Saturdays included. 3 strike system: a missed day with no message ahead is a strike, and repeating a mistake after feedback is a strike. You need a computer that can export 7 videos a day and reliable internet.
HOW TO APPLY
Start your message with the word "OMNI". Then:
1. Link 2 to 3 AI talking-avatar videos you made yourself WITH OMNI. For each one: which client or brand, and what you personally did.
2. How long have you been using Omni, and roughly how many finished videos have you delivered with it?
3. Tell me about one Omni clip that got blocked, cut off a word, or drifted off the start frame, and exactly how you fixed it.
4. Show one character you kept consistent across 2 or more videos (links).
5. Honestly: how many 1 to 2 minute videos can you finish per day at my example's quality?
6. Your working hours and which days you commit to.
Applications without Omni example videos get no reply.
WHAT YOU GET FROM ME
- Scripts, reference videos, character images and start frames for every video, sent the day before
- My exact process and prompt library, plus Loom feedback
- Accounts and credits provided
- Long-term stable work, a raise as you get faster, and a monthly performance bonus