Internal Tools¶
This document describes the internal applications that support the main Sartiq platform. These tools are used by the Sartiq team to create and manage assets that power the platform.
Overview¶
flowchart TB
subgraph Internal [Internal Tools]
Identiq[Identiq]
Movieq[Movieq]
end
subgraph Platform [Sartiq Platform]
Backend[Backend API]
Compute[Compute Server]
end
Identiq -->|Subjects| Backend
Movieq -->|Videos| Backend
Movieq --> Compute
The internal tools ecosystem consists of:
| Tool | Purpose | Status |
|---|---|---|
| Identiq | Subject (AI identity) creation | Active |
| Movieq | Video generation from images | Integration in progress |
Identiq¶
Identiq is the internal platform for creating and managing Subjects - the crafted AI identities used as models in generated imagery.
What Identiq Does¶
Identiq enables the team to:
- Create new subjects from reference images
- Refine subject identities through iterative prompt engineering
- Generate consistent appearances across different poses and settings
- Manage the subject library used by the main platform
Why Subjects Matter¶
Subjects are essential for maintaining consistency in generated imagery. A single subject can appear across hundreds of generated images while maintaining recognizable identity - similar to how a real model would appear consistent across a photo shoot.
Subject Creation Workflow¶
flowchart TB
subgraph Input["1. Input Collection"]
Ref[Reference Images]
Brief[Creative Brief]
end
subgraph Creation["2. Identity Creation"]
Prompt[Prompt Engineering]
Gen[Initial Generation]
Review[Review & Iterate]
end
subgraph Refinement["3. Refinement"]
Variations[Generate Variations]
Select[Select Best Results]
BaseImg[Create Base Images]
end
subgraph Output["4. Output"]
Subject[Subject Record]
Platform[Available in Platform]
end
Ref --> Prompt
Brief --> Prompt
Prompt --> Gen
Gen --> Review
Review -->|iterate| Prompt
Review -->|approved| Variations
Variations --> Select
Select --> BaseImg
BaseImg --> Subject
Subject --> Platform
Creative Process¶
The subject creation process involves several stages:
1. Reference Collection¶
- Gather inspiration images that define the desired look
- Document demographic characteristics and style notes
- Define the subject's "personality" and typical poses
2. Prompt Engineering¶
- Craft detailed prompts that capture the subject's appearance
- Iterate on prompt variations to find the right balance
- Test prompts across different scenarios
3. Generation & Refinement¶
- Generate initial identity images
- Review for consistency and quality
- Refine prompts based on results
- Create multiple variations to verify consistency
4. Base Image Creation¶
- Generate a pack of base images showing the subject in various shot types
- Test base images in various generation scenarios
- Validate consistency across different styles and poses
Subject Properties¶
| Property | Description |
|---|---|
| Name | Display name for the subject |
| Gender | Male, Female, or Other |
| Age Class | Kid, Teen, Young, Middle Age, Elder |
| Cover Image | Primary display image |
| Base Images | Pack of images showing the subject in various shot types |
| Description | Optional text describing the subject |
Integration with Main Platform¶
Subjects created in Identiq are synced to the main platform's database and become available for selection in shooting configurations.
sequenceDiagram
participant Team as Sartiq Team
participant Identiq
participant DB as Platform DB
participant Platform as Main Platform
Team->>Identiq: Create subject
Identiq->>Identiq: Generate & refine
Team->>Identiq: Approve subject
Identiq->>DB: Sync subject record
DB-->>Platform: Subject available
Note over Platform: Subject now available
Future: Customer Subject Requests¶
A planned feature will allow customers to request custom subjects through the main platform. These requests would be fulfilled through Identiq:
- Customer submits subject request with requirements
- Sartiq team receives request in Identiq
- Team creates subject following customer specs
- Subject is made available exclusively to that customer
Movieq¶
Movieq is the internal video generation platform, designed to create video content from images and prompts.
What Movieq Does¶
Movieq enables video generation capabilities:
- Image-to-video conversion - Animate static images
- Motion control - Define camera movements and subject motion
- Video editing - Combine and enhance video clips
Current Status¶
Movieq is currently a standalone internal tool. Integration with the main Sartiq platform is planned, which will allow:
- Video shot types in shooting configurations
- Animated product showcases
- Dynamic lifestyle content
Video Generation Workflow¶
flowchart LR
subgraph Input
Image[Source Image]
Motion[Motion Parameters]
Prompt[Animation Prompt]
end
subgraph Process
Generate[Video Generation]
Preview[Preview & Review]
end
subgraph Output
Video[Final Video]
end
Image --> Generate
Motion --> Generate
Prompt --> Generate
Generate --> Preview
Preview -->|revise| Generate
Preview -->|approve| Video
Capabilities¶
| Feature | Description |
|---|---|
| Image-to-Video | Convert still images to video with motion |
| Camera Motion | Pan, zoom, dolly movements |
| Subject Animation | Natural movement of people and objects |
| Duration Control | Generate clips of various lengths |
| Style Transfer | Apply visual styles to videos |
Motion Parameters¶
Video generation can be controlled through parameters:
# Example motion configuration
motion_config = {
"camera": {
"movement": "slow_zoom_in",
"start_frame": {"x": 0, "y": 0, "zoom": 1.0},
"end_frame": {"x": 0, "y": 0, "zoom": 1.2},
},
"subject": {
"motion": "subtle_movement",
"intensity": 0.3,
},
"duration": 3.0, # seconds
"fps": 24,
}
Integration Plans¶
When integrated with the main platform, Movieq will:
- Add video shot types - Full body video, product turnaround, etc.
- Extend generation workflows - Include video tasks in DAG
- Enable video review - Gallery support for video content
- Support video export - Deliver video alongside images
flowchart TB
subgraph Current["Current Flow"]
SL1[Shooting Look] --> Shot1[Image Shots]
Shot1 --> Gen1[Image Generation]
end
subgraph Future["Future Flow"]
SL2[Shooting Look] --> Shot2[Image Shots]
SL2 --> Shot3[Video Shots]
Shot2 --> Gen2[Image Generation]
Shot3 --> Gen3[Video Generation]
end
Tool Architecture¶
Shared Infrastructure¶
Both internal tools share infrastructure with the main platform:
| Component | Shared? | Notes |
|---|---|---|
| Database | Partial | Separate DBs, synced entities |
| Compute Server | Yes | Same AI orchestration layer |
| Storage | Yes | Same object storage |
| Auth | Separate | Internal SSO |
Data Flow¶
flowchart TB
subgraph Identiq
ID_UI[Identiq UI]
ID_API[Identiq API]
end
subgraph Movieq
MQ_UI[Movieq UI]
MQ_API[Movieq API]
end
subgraph Shared["Shared Services"]
Compute[Compute Server]
Storage[(Object Storage)]
end
subgraph Platform["Main Platform"]
Backend[Backend API]
DB[(Platform DB)]
end
ID_API --> Compute
ID_API --> Storage
ID_API -->|sync subjects| DB
MQ_API --> Compute
MQ_API --> Storage
MQ_API -.->|future| Backend
Access & Security¶
Who Uses These Tools¶
| Tool | Users |
|---|---|
| Identiq | Creative team, AI engineers |
| Movieq | Creative team, content producers |
Authentication¶
Internal tools use separate authentication:
- SSO Integration - Company identity provider
- Role-Based Access - Different permissions per tool
- Audit Logging - Track all operations
Related Documentation¶
- Domain Concepts - Understanding Subjects
- Compute Server - Shared AI infrastructure
- Generation Pipeline - How subjects are used