How the autopilot works.
PhantomLab doesn't generate content in one-shot batches. Once you set it up, it runs — planning, generating, and surfacing stories and feed posts for approval on an ongoing cadence. Here's exactly how.
The loop.
After initial setup, PhantomLab runs a continuous loop: the system plans upcoming content, generates it in advance, surfaces each piece to you for approval, and publishes on approval. Your only ongoing involvement is a single tap — approve, regenerate, or skip.
Plan
The autopilot decides what the persona should post next — considering cadence, recent content, seasonal context, and the persona's established style.
Generate
Scene, wardrobe, composition, lighting, and pose are engineered. The image renders through the full six-stage pipeline.
Notify
The finished post appears in your notification channel with a full preview. Ready for a yes/no decision.
Approve
You tap approve — or reject and request a regeneration. One-tap interaction, nothing more.
Publish
Approved content posts to the connected platform. The system logs the outcome and loops back to Plan.
The autopilot respects the posting rhythm you set for the persona — daily, multiple times a week, whatever matches the persona's established pattern. It doesn't flood. It doesn't lapse.
Inside each generation.
Every time the autopilot produces a post, the image runs through a six-stage pipeline. This is the engineering that ensures character identity stays locked, scenes stay coherent, and output quality stays at commercial grade.
Analyze
Extract facial features, body type, skin tone from reference image.
Plan
LLM designs each shot: context, lighting, wardrobe, camera angle, mood — calibrated for commercial image quality.
Engineer
Multi-pass prompt construction with realism constraints and content-safety rewording.
Evaluate
Quality analysis checks naturalness, composition, framing — ensures output meets commercial standards.
Refine
Output description feeds back into the prompt pipeline for iterative improvement.
Deliver
Consistent, commercial-quality results with character identity locked across all scenes.
What the autopilot can produce.
These are the creative dimensions the autopilot draws from when planning content. It selects scenes, styles, and formats based on the persona's profile and posting history — this is the palette it works with.
environments
Lifestyle settings
Four natural lifestyle environments — cafés, streets, interiors, outdoors — the autopilot uses for editorial and campaign content. Each environment is framed with natural composition and professional lighting, so posts feel like real moments rather than staged shoots.
environments
Professional settings
Ten profession-specific workplace and studio environments: fashion studios, corporate offices, creative agencies, labs, and more. The autopilot selects the setting that matches the persona's declared profession, so content stays narratively consistent with who the character is.
formats
Stories and feed posts
The autopilot plans in both formats — vertical stories for daily presence and feed posts for anchor content. It alternates between them based on posting cadence, so the persona's timeline reads like a real account rather than a gallery of one format.
styling options
Wardrobe and styling range
Every planned shoot includes wardrobe, accessories, and styling decisions calibrated to the persona's established aesthetic. The autopilot varies outfits across posts the way a real person's feed varies — seasonal shifts, context-appropriate dress, style continuity without repetition.
identity
Locked character identity
Every piece of content the autopilot produces preserves the same face, features, body, and styling logic from the original reference. Across hundreds of posts and months of output, the persona remains recognizably the same person.
max length
Motion and video
The autopilot can escalate any still into a 3–15 second video clip using Kling 3.0 Pro. Natural movement, expression, and environmental motion — for the moments when a persona's feed calls for more than a photograph.
One-time setup.
Before the autopilot runs, you configure it once. Four steps.
Upload your reference
Drop in a single reference image of your virtual character. This is the identity anchor — the autopilot preserves this face across every post it generates.
Configure scenes
Select Lifestyle mode for 4 natural environment settings, Professional mode for 10 profession-specific workplace scenes, or both. This defines the creative palette the autopilot draws from.
System plans your content
The LLM analyzes your character and plans photorealistic content for each scene. Wardrobe, lighting, composition, pose — all calibrated before the first image renders.
Autopilot engages
From here, the autopilot takes over. Content renders in parallel, posts surface for your approval, and the system maintains your persona's posting cadence without further input.
Common questions.
What happens if I don't approve a post?
Nothing publishes without your approval. Rejected posts can be regenerated with different parameters or skipped entirely. The autopilot moves to the next planned piece.
Can I pause the autopilot?
Yes, at any time. Pause and resume whenever you want — the system picks up where it left off.
Is the content safe for commercial use?
Yes. PhantomLab is built for commercial and editorial use. The pipeline includes content-safety rewording at the prompt engineering stage. See our Acceptable Use Policy for full details.
Ready to set yours up?
The autopilot is one upload and a short persona brief away. Everything else runs in the background.
Start Creating