Running locally? Start the backend and app first (e.g. run START_HERE.bat from the repo root), then click Open LexAI App. See docs/integrated_local_run_guide.md in the repo for the full steps.

Provenance for creative work

Show how your studio and AI built the final image.

LexAI tells the story behind each deliverable — which parts came from your team, which from AI, and how they came together into the final design.

Open LexAI App Request a demo Tracker setup & download

This site is the public website. The LexAI Provenance product app lives in the frontend/ folder of this repo. For local run steps, see the note above.

The problem: AI makes provenance hard

Clients ask “how much is AI?”

Teams use Gemini, ChatGPT, and other models alongside Photoshop or Figma. When a client or platform asks how much of the work was AI-generated, most teams answer from memory.

Legal and platforms need proof

Policies increasingly require teams to explain which tools were used and who contributed what. Without a provenance layer, it's hard to provide anything beyond screenshots and narrative.

LexAI Provenance — four product views

LexAI Proof

A summary page that shows human vs AI contribution, key editing actions, evidence strength, and any risk notes for a single result.

LexAI Overlay

Overlay human and AI contribution directly on top of the final image. Click units to see where they apply on the canvas.

LexAI Split

Break down contribution scores by unit to show how much of the final result was human-led vs AI-assisted, overall and per slice.

LexAI Report

Generate a certificate-style report that you can print, save as PDF, or share via link for clients, legal, or submissions.

Seven tools, one contribution story

LexAI is built around a realistic seven-tool stack, so design teams can start with the tools they already use.

Desktop & design
Photoshop Paint Figma Canva
AI models
ChatGPT Gemini nanobanana

Desktop tracking

Lightweight agent watches saves and exports from Photoshop and Paint, linked to a project.

Browser observed

Best-effort capture of AI and design activity in the browser for tools like Gemini or Figma.

Export linked

Exports from Figma or Canva are recorded as key events for the final deliverable.

User confirmation

Designers can confirm or correct which sessions and outputs belong to a project before proof.

In the LexAI app, these tools and capture styles feed into one contribution model, so Proof, Overlay, Split, and Report all tell the same story for a single result.

Built for collaborative creative teams

LexAI is designed for design companies, content studios, and in-house creative teams who work across many tools and contributors on the same project.

Who uses LexAI

  • Brand and design studios running multi-tool campaigns.
  • Content and marketing teams combining AI and human workflows.
  • In-house creative teams collaborating with external partners.
  • Production teams preparing assets for client or platform review.

Typical workflow in LexAI

  • Track a project across tools like Photoshop, Figma, Canva, Gemini, and ChatGPT.
  • Let multiple contributors work in one shared project with tracked activity.
  • Review contribution in LexAI Proof, Overlay, and Split for a single final result.
  • Generate a LexAI Report to share with clients, legal, or platforms as part of submission or review.

This section focuses on practical business use: multi-source creative work inside a single project, clear attribution review, and shareable proof for external stakeholders.

Implementation & patent-linked documentation

The sections below summarize how the LexAI Provenance app is implemented and prepared around final-result analysis, multi-source attribution, and rights-proof / settlement concepts. They are written for product, founder presentation, and evidence use.

One story from AI prompt to final export

The core LexAI story is simple: Gemini → nanobanana → Photoshop, all visible in one timeline and one report.

  • Capture AI prompts and responses through connectors or scripts.
  • Link desktop edits and exports through a light-weight agent.
  • Turn everything into a clear, shareable proof for a single deliverable.

Try LexAI with your own stack

Start with the same seven-tool story we use for demos, then layer in your own tools and workflows.

Request early access Tracker setup & download

Download the desktop tracker ZIP, extract it, set your project id, and run — events show in the LexAI app Project Home when the backend is reachable.

Desktop tracker (agent) for real tool activity

LexAI uses a lightweight Windows agent to capture saves, exports, and app activity from tools like Photoshop. Use the download below, then point it at your project in the LexAI app.

Download desktop tracker (ZIP)

How to run it (Windows)

  • Unzip LexAI-Tracker-Windows.zip. You should see LexAI-Tracker.exe, .env.example, SETUP.txt, and README.md (no Python required).
  • Run LexAI-Tracker.exe once: it creates .env automatically if missing. Set LEXAI_AGENT_PROJECT_ID to the number from your LexAI Hub URL (?project_id=...).
  • Run LexAI-Tracker.exe again after editing .env. Leave the console window open while you work; events go to LexAI when the backend is reachable.
  • Advanced: use lexai-desktop-tracker.zip from the same releases folder for Python source + pip install -r requirements.txt and python main.py.

What you need in the app

  • Open LexAI Hub and use the same project_id you configured for the agent.
  • Full details: see README.md inside the downloaded agent folder, or the agent/README.md file in the GitHub repository.
  • A signed installer may come later; the ZIP is the supported pilot path today.

1. Implementation flow — from project to proof

LexAI Provenance is being developed together with Korean patent applications including “최종 결과물 기준 이벤트 연계 분석을 통한 창작 과정 이벤트 처리 방법 및 시스템” and related inventions around multi-source contribution attribution and rights-proof preparation. The flow below summarizes how the product reflects these ideas in practice.

  1. Project creation (implemented)
    A project is created in LexAI. Project Home becomes the place where activity, tools, and candidate outputs are grouped for that project.
  2. Multi-tool creative activity (implemented)
    Designers work across tools such as Photoshop, Figma, Canva, Paint, Gemini, ChatGPT, and nanobanana. These tools are visible as “Tools we're tracking” on Project Home.
  3. Event capture (implemented)
    A desktop agent and connectors capture app activity, file saves/exports, and AI sessions as events linked to the project. Recent activity is shown in Project Home, reflecting the multi-source event stream (multi-source collaboration concept, 3rd patent).
  4. Final result selection (implemented)
    One output (artifact + version) is chosen as the final result for review. Hub, Viewer, Proof, Overlay, Split, and Report all share this selection.
  5. Final-result-centered event linking (implemented)
    For the selected result, backend logic links back to the events that contributed to that specific artifact/version. This is surfaced in LexAI Proof, Overlay, and Split as contribution units and regions (final-result-based linkage concept, 1st patent).
  6. Contribution attribution across tools and participants (implemented)
    LexAI Proof, Overlay, and Split show how contribution is attributed across human vs AI and across different tools and steps. This reflects multi-source collaborative contribution analysis (3rd patent) using concrete scores, regions, and timelines.
  7. Proof / report / certificate generation (implemented)
    LexAI Report generates a print-ready report that summarizes project, result, tools used, contribution shares, evidence level, and risk notes. This is used for client review and internal/legal review (certificate-style proof for the final result, 1st and 3rd patents).
  8. Future rights-proof / settlement extension (planned)
    The same final-result-centered contribution data can be extended to build rights-proof packages and atomic settlement logic for creators and partners (4th patent direction). This is a documented extension of the current product, not yet exposed as a full feature.

This implementation flow is product-facing documentation only and does not describe legal claims. It is intended to show that the patented concepts are being implemented and prepared in a real system as part of LexAI's product work.

2. Current and planned LexAI capabilities

The table below summarizes concrete LexAI features that relate to final-result-based event analysis, multi-source contribution attribution, and future rights-proof / settlement preparation.

Implemented in the product today

  • Final-result-centered event analysis
    Artifact + version selection in the app; backend links events to that final result and surfaces them in LexAI Proof, Overlay, and Split.
  • Tool-linked creative activity tracking
    Desktop agent and connectors record activity from Photoshop, Figma, Canva, Paint, Gemini, ChatGPT, and nanobanana into project-scoped event streams.
  • Contribution review screens
    LexAI Proof, Overlay, and Split provide contribution units, regions, and scores for the final result, enabling human vs AI review by design teams and reviewers.
  • Multi-source support across tools
    Events from multiple tools and models are normalized into a shared contribution model so that all Proof / Overlay / Split / Report views stay consistent for the same result.
  • Split / contribution allocation structure
    LexAI Split aggregates contribution scores into overall human vs AI share and per-unit allocation, reflecting how different steps and tools contributed.
  • Report / certificate outputs
    LexAI Report generates a printable, shareable report that captures project context, tools used, contribution shares, evidence level, and risk notes for the selected result.

Prepared / expanding capabilities

  • Deeper multi-source collaboration modeling
    Extending contribution units and event types to represent more complex collaboration patterns across tools, roles, and sessions based on the same event-linkage structure.
  • Rights-proof package structure
    Designing a package format that can bundle final-result-centered contribution data, evidence references, and policy metadata for downstream use (e.g. platforms, agencies, or partners).
  • Atomic settlement preparation
    Exploring how contribution allocations and rights-proof packages can be used as inputs to future settlement or compensation logic, while reusing the same project and artifact/version structure.
  • Additional connector coverage
    Preparing connectors and scripts for more direct event ingestion from design and AI tools, beyond the core seven-tool story, using the same Provenance model.

These capabilities are described as product features and roadmap items only. They are intended to provide clear, screenshot-friendly evidence of how LexAI operationalizes and prepares the concepts in the related patent applications.

3. How LexAI screens map to these concepts

The four main LexAI screens below are concrete UI surfaces that implement final-result-based analysis, multi-source contribution attribution, and (for the report) rights-proof preparation.

LexAI Proof

Final-result-based summary view for a single artifact and version. Shows human vs AI contribution, tools used, evidence level, and risk notes for that result (1st and 3rd patent-linked concepts).

LexAI Overlay

Visual layer on top of the final image where each contribution unit is mapped to regions. Connects event-linked units to specific parts of the final result and makes multi-source attribution visible on the canvas (1st and 3rd patent-linked concepts).

LexAI Split

Table and summary that allocate contribution scores across human vs AI and per unit. Uses final-result event linkage and multi-source scores to show how different tools and participants contributed (1st and 3rd patent-linked concepts).

LexAI Report

Printable, shareable report for the selected result, combining project context, tools, contribution shares, and evidence notes into one document. Serves as a practical proof/certificate output today and a base for future rights-proof and settlement packages (1st, 3rd, and 4th patent-linked concepts).

This screen mapping is written for product and evidence purposes only. It is intended to make it easy to capture screenshots or PDFs that show how LexAI operationalizes each concept in a real interface.

4. Certificate outputs and future rights-proof extensions

LexAI already generates certificate-style reports for individual results and is being prepared for use as a rights-proof package and, later, as an input to settlement or licensing flows.

Current capabilities

  • Certificate-style report for each result
    LexAI Report produces a print-ready, shareable document with project context, selected artifact and version, tools used, contribution shares, evidence level, and risk notes for that single result.
  • Content-linked proof outputs
    Reports and views in LexAI are always anchored to a specific project and final deliverable, so proof is explicitly tied to the content under review rather than being generic.
  • Consistent data across views
    Proof, Overlay, Split, and Report all read from the same contribution data for a result, so the certificate reflects the same attribution and evidence that the internal team sees in the app.

Prepared / future extensions

  • Rights-proof package structure
    LexAI's contribution and event model is being organized so that the report, underlying scores, and key events can be bundled as a rights-proof package for external platforms, agencies, or partners.
  • Settlement and allocation inputs
    Contribution shares and capsule-level attribution are being designed to serve as inputs to future settlement, licensing, or allocation logic, while keeping the core Provenance data model unchanged.
  • Licensing and platform workflows
    Product planning includes flows where a LexAI report or rights-proof package can be submitted along with content to marketplaces or compliance systems, without changing the creative tools teams already use.

This section describes product behavior and planned extensions only. It is intended to support evidence that LexAI has a real certificate/report layer today and is being actively prepared for rights-proof and settlement-related use cases, without claiming that full legal settlement logic is already live.

Implementation evidence highlights

The points below summarize concrete LexAI Provenance behaviors that can be used as evidence of implementation and preparation for the related patent concepts.

  • Project-based tracking — events from desktop tools, AI models, and exports are recorded into project-scoped streams visible on Project Home.
  • Final-result-centered analysis — a single artifact and version is selected as the final result, and LexAI Proof, Overlay, Split, and Report all operate on that specific deliverable.
  • Multi-source contribution attribution — contribution units and scores combine activity from multiple tools and participants into one coherent view for the result.
  • Proof / report / certificate outputs — LexAI Report generates a printable, shareable document with project context, tools, contribution shares, and evidence/risk notes.
  • Rights-proof / settlement preparation — the contribution and report structure is being organized so it can serve as a rights-proof package and, later, as an input to settlement or licensing logic.

This highlight block is written purely as a product summary and is intended to be easy to capture as a single screenshot or PDF section when assembling supporting material.

5. Related patent applications and product context

The LexAI Provenance product is being developed together with several Korean patent applications. The table below summarizes key applications and how they relate to the current implementation work.

Patent concept Title (Korean) Application no. LexAI Provenance context
First patent 최종 결과물 기준 이벤트 연계 분석을 통한 창작 과정 이벤트 처리 방법 및 시스템 (to be filled with official number) Implemented today as final-result-based event linkage and analysis for a single artifact and version, surfaced in Project Home, LexAI Proof, LexAI Overlay, LexAI Split, and LexAI Report.
Third patent (multi-source collaborative contribution attribution concept) (to be filled with official number) Reflected in multi-tool, multi-contributor event capture and attribution across tools and participants, as shown in the seven-tool support, project activity streams, and contribution views (Proof, Overlay, Split, Report).
Fourth patent (rights-proof package / atomic settlement preparation concept) (to be filled with official number) Reflected in the current certificate-style LexAI Report and in the prepared direction for rights-proof packages and future settlement/licensing inputs based on final-result-centered contribution data.

This block is for product and documentation purposes only. It is intended as structured, screenshot-friendly reference material and does not state or interpret legal claims.

Patent concepts & LexAI product mapping

The table below summarizes, at a glance, how each patent concept relates to specific LexAI Provenance functions and screens.

Patent concept Product function Related screen / workflow Implementation / preparation context
First patent
Final-result analysis		 Final-result-based event analysis for a single artifact and version. Project Home selection, LexAI Proof, LexAI Overlay, LexAI Split, LexAI Report. Implemented: events are linked to the selected result and analyzed across these screens to show human vs AI contribution, regions, and shares.
Third patent
Multi-source attribution		 Multi-tool, multi-contributor contribution attribution in one model. Seven-tool tracking, project activity streams, contribution units and scores in LexAI Proof, Overlay, Split, and Report. Implemented: events from desktop tools and AI models are normalized into shared contribution units and scores so teams can review how different sources contributed to one result.
Fourth patent
Rights-proof / settlement		 Certificate-style outputs and preparation for rights-proof packages and settlement inputs. LexAI Report; future rights-proof package and platform/settlement workflows. Implemented: LexAI Report as a certificate-style output for each result. Prepared: organizing report and contribution data so it can be bundled as rights-proof and used as an input to later settlement/licensing logic.

This summary is written from a product perspective and is intended to be easy to capture as a single table when compiling supporting evidence. It does not attempt to restate or interpret any legal claims.