My AI Avatar Clone is So Realistic It Replaced Me
An 18-minute walkthrough of building a photorealistic AI video clone from a selfie, an audio clip, and a structured timeline prompt.
May 19thA free, vibe-coded dashboard that keeps your AI influencer roster on-model across images, outfits, and UGC videos — no repeated prompting required.
Managing multiple AI influencer personas across platforms is a prompt-juggling nightmare that a single locally-hosted dashboard eliminates by storing each character identity, outfit history, and brand deal assets in one persistent profile.
Running five AI influencer pages at scale means constantly switching between Higgsfield, GPT Image 2, and Nano Banana and re-prompting each character from scratch every session. The dashboard presented here solves that by storing each influencer as a persistent profile — physical traits, backstory, personality, saved wardrobe variants, and brand deal product sheets — and routing all generation through a single local UI. Setup takes four steps (Antigravity IDE, Claude Code extension, GitHub download, npm install) and the optional Claude API connection improves prompt quality without much spend. The full flow from reference image to character sheet to outfit to scripted UGC video takes under ten minutes once configured.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →
Five AI influencer pages doing millions of monthly impressions; the multi-platform prompt workflow is the bottleneck; introducing the dashboard.

Nine pre-built influencers shown; character sheet generation, outfit swaps, product placement, and video output demonstrated.

Side-by-side phone mockups: hiking/hydration UGC and Keychron keyboard UGC played simultaneously.

Four steps: download Antigravity IDE, install Claude Code extension, download GitHub ZIP, run npm install. Launch on localhost.

Connect Higgsfield API via OAuth; add optional Claude API key for smarter prompt generation.

Name, gender, age, niche (finance), reference photo, backstory, personality slider, physical traits, aesthetic vibe (cottagecore).

Select GPT Image 2; generate 3 looks; pick canonical look; generate character sheet, close-ups, feature sheet; fill profile metadata.

Generate and name outfit variants; upload room for world-building; add brand products and generate multi-angle product sheets.

Combine location, time of day, pose, expression, outfit, and props in any combination.

Script input, product add, shot type, voice/accent settings, resolution. Luna lamp ad and Marcus Keychron ad demonstrated.

Mountain lifestyle images plus Hydro Flask hiking vlog UGC video.

Vibe-coded in two weeks with no prior coding; free because unfinished; assets in Skool (NextGen AI); Higgsfield MCP as next frontier.
Visual consistency across AI influencer content breaks down the moment you switch platforms or sessions — the fix is storing the full identity as a persistent profile, not re-prompting from scratch each time.
“I figured out the easiest way to create AI influencers, and this might be the future of how you start using AI.”
“I'm giving you this entire tool for completely free because I customized it and vibe coded myself, and I see no point in selling something like this if it's not a finished product.”
“I had one problem and I vibe coded my own solution. I don't have any coding experience. I just got started with this, like, a few weeks ago.”
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
Five AI influencer pages. Millions of monthly impressions. And a workflow so fragmented — Nano Banana for one thing, GPT Image 2 for another, Higgsfield for video — that the creator vibe-coded his way out of it in two weeks. The result is a free, locally-run dashboard that keeps every persona on-model from first character sheet to final UGC ad.
The three-asset foundation that locks visual identity before any downstream generation. Everything else references these three files.
Three-layer identity scaffold. The reference image anchors looks; the physical traits resolve ambiguity; the aesthetic vibe shapes all outfit and setting prompts automatically.
“Make sure that you download these assets from my school community. I will also provide a bit more of a guide on how to install it.”
Soft, no hard sell. Points to free Skool community (NextGen AI) shown on screen. Low friction.
00:00
00:12
00:29
00:43
00:50
01:02
01:09
01:25
01:38
01:51
02:00
02:11
02:17
02:34
02:45
02:57
03:08
03:20
03:33
03:44
03:52
04:05
04:19
04:26
04:40
04:47
05:03
05:14
05:26
05:35
05:48
06:00
06:11
06:23
06:34
06:46
06:51
07:10
07:20
07:31
07:43
07:54
08:03
08:14
08:33
08:37
08:47
09:02
09:09
09:26
09:41
09:48
10:00
10:17
10:20
10:39
10:41
10:57
11:09
11:20
11:31
11:48
11:51
12:06
12:15
12:28
12:36
12:47
13:02
13:14
13:21
13:32
13:44
14:03
14:12
14:23
14:32
14:46
14:56
15:09An 18-minute walkthrough of building a photorealistic AI video clone from a selfie, an audio clip, and a structured timeline prompt.
May 19thA 22-minute tutorial that builds an end-to-end AI creative agency stack — product images, UGC ads, cinematic commercials, and a self-improving Notion back end — using four tools wired together through Claude.
June 8thA 64-minute masterclass where a Codex true believer converts a Claude Code skeptic, live, on camera.
April 27thHow 10 hours of SKILL.md investment turns a Claude Code session into a one-click editor for 18 shorts at once.
June 5thA working agentic workflow that handles rough cuts, overlays, sound effects, memes, zooms, short-form repurposing, and thumbnail selection — all from Claude.
May 18thA 26-minute listicle-tutorial demoing 17 Claude Skills across the full content creation stack — from goal alignment to newsletter drafting.
June 1st