PromptArt

A graph-native system for crowdsourced GenAI content transformation, attribution, and value circulation.

Motivation

Digital content ecosystems are structurally fragmented:

• generation is powerful but often isolated,
• lineage is weak across derivative outputs,
• value distribution rarely reflects true contribution paths.

PromptArt addresses this gap by treating generation, provenance, and settlement as one integrated graph process rather than separate products.

Vision

PromptArt advances a Crowdsourced Content Transformation Graph (CCTG) in which:

• source content, transformation operators, and generated artifacts coexist in one evolving network,
• every transformation becomes a reusable graph edge,
• attribution and economic flow are computed over the same graph paths that produce content.

The long-term outcome is a creative infrastructure where personalization scales without disconnecting provenance and contributor upside.

Core Technology

1. Graph-Native Transformation Substrate

The platform models content evolution as a directed graph:

• roots: ingested sources,
• operators: atomic and composed transformers,
• outputs: multimodal artifacts,
• return paths: attribution and settlement edges.

2. Multimodal GenAI Operator Layer

Transformation edges are powered by language, image, and audio adapters, enabling:

• text rewriting/synthesis,
• text-to-image generation,
• speech-to-text and text-to-speech conversion,
• mixed-modality transformation chains.

3. Attribution and Settlement Logic

Consumption and reuse events are tied to graph lineage, enabling:

• contribution-aware allocation,
• rights-gated access across derived artifacts,
• settlement updates aligned to transformation ancestry.

Impact Across User Groups

Consumers

• Receive richer, personalized outputs from the same source corpus.
• Benefit from continuously improving transformation pathways.

Creative Artists

• Publish transformation operators as reusable graph components.
• Capture downstream value when their operators are reused.

Content Providers

• Contribute source nodes that remain traceable across derivations.
• Participate in value flow as content propagates through the graph.

Model Providers

• Supply core inference capabilities in language/image/audio layers.
• Gain participation in transformation-driven economic activity.

Implementation and Architecture

The current repository implements an AWS-based asynchronous pipeline:

API Gateway -> API Lambda -> Task Queue (SQS) -> Worker Dispatch -> Transformer Runtime
                                  |                                      |
                                  v                                      v
                          Graph/Doc/Right State                  Multimodal Model Adapters
                                  |                                      |
                                  +---------- Attribution + Settlement ---+

Implemented Architecture Components

• API orchestration: aws/projects/prompt-art-api/lambda_function_api.py
• Queue workers: aws/projects/prompt-art-sqs/lambda_function_sqs.py and Docker events manager
• Task engine: aws/core/paTasks.py
• Dispatch + charging flow: aws/core/paDispatch.py
• Transformer composition/runtime: aws/core/paTransform.py, aws/core/paTransformSrv.py
• Graph/feed management: aws/core/paGraph.py
• Document rights and access control: aws/core/paDocs.py
• Wallet/transfer primitives: aws/core/paUsers.py
• Media persistence layer: aws/core/paMedia.py
• Adapter registry (language/image/audio/source): aws/core/paAtomic.py

Current Technical State

• End-to-end async generation exists.
• Graph-linked feed and transformer orchestration exists.
• Rights checks and token charging exist.
• Foundation for attribution-aware settlement exists and is extensible toward richer ledger-grade accounting.

---

PromptArt positions generative media as a graph systems problem: composition, provenance, and value distribution become first-class properties of the same computational structure.