Zing Forum

Reading

Franklin Canvas: A Node Editor-Based AI Media Creation Platform with Real-Time USDC Payments

An infinite canvas node editor supporting image, video, and music generation, enabling real-time USDC payments via the x402 protocol—no subscriptions, pay-per-use

AI生成节点编辑器x402支付USDC图像生成视频生成音乐生成Web3按次付费BlockRun
Published 2026-05-31 10:45Recent activity 2026-05-31 11:23Estimated read 9 min
Franklin Canvas: A Node Editor-Based AI Media Creation Platform with Real-Time USDC Payments
1

Section 01

Franklin Canvas: Node-based AI Media Creation Platform with USDC Real-time Payment

Core Overview

Franklin Canvas is an AI media creation platform developed by BlockRunAI, released on GitHub on May 31, 2026 (source: GitHub repo). It combines an infinite canvas node editor with multi-modal AI generation (image, video, music, text) and innovative Web3 payment via x402 protocol for real-time USDC settlement. Key features include pay-per-use (no subscription), support for Base and Solana chains, and flexible workflow orchestration.

Key Highlights

  • Node-based visual workflow for complex AI generation tasks
  • Multi-modal support: image (text2img/img2img), video (5-30s), music (3min), text
  • Real-time USDC payment via x402 protocol (no subscription fees)
  • Open-source project by BlockRunAI
2

Section 02

Background: Payment Dilemmas in AI Media Creation

With the rapid development of AI generation technology, most platforms adopt subscription models, which have two main issues:

  1. Fixed cost waste: Users pay monthly/annual fees regardless of actual usage frequency, unreasonable for creators with fluctuating needs.
  2. Limited model flexibility: Subscription plans often restrict users to specific vendor ecosystems, making it hard to choose the best model for different tasks.

These problems create barriers for creators seeking flexible, cost-effective AI tools.

3

Section 03

Core Solution & Key Features

Franklin Canvas addresses the above issues with the following core features:

1. Infinite Canvas & Node Editing

Users can drag nodes, connect workflows on an infinite canvas, with support for pan/zoom, mini-map navigation, and Figma-style group dragging.

2. Multi-modal Generation Nodes

  • Image Node: Text-to-image (text2img) and image-to-image (img2img) with multiple models.
  • Video Node: 5-30s video clips, supporting seed images and audio options.
  • Music Node: ~3min music tracks with adaptive or custom lyrics.
  • Text Node: Chat model for scripts, copywriting, and brand messaging.

####3. Pay-per-use Without Subscription Each generation call is settled in real-time via x402 protocol using USDC, supporting Base and Solana chains. Users only need to deposit USDC in their wallet—no subscription fees.

4

Section 04

Technical Architecture & x402 Payment Flow

Frontend Stack

  • Vite + React 19 + TypeScript
  • @xyflow/react (node editor canvas)
  • Zustand (lightweight state management)
  • Vanilla CSS variables (3 themes: Dark/Gold/Light)

Backend Architecture

  • Pure Node.js (no framework) with @blockrun/llm SDK for wallet management and x402 payment signing.
  • Flow: Browser → Backend → BlockRun Gateway → AI models.

x402 Payment Process

  1. Backend returns HTTP 402 response with payment challenge when a generation request is made.
  2. Client signs the response using their wallet.
  3. Backend validates the signature and executes the generation.
  4. Result is returned to the client.

Video generation uses manual polling (instead of auto-polling) to handle potential 402 re-signing and supports up to 20min wait for slow models like Seedance.

5

Section 05

Rich Creation Features

Context Prompt Bar

Select a node to input prompts; attach reference images from the canvas for img2img or img2video.

In-canvas Image Editing

  • Paid operations: Outpaint/Enhance/Cutout/Upscale (via BlockRun gateway).
  • Free operations: Split into 2×2/3×3 tiles, annotate (hand-drawn), cutout to transparent PNG.

Prompt Library

800+ curated prompts from BlockRunAI's open-source case library, filterable by workflow type (text2img, img2img, etc.) and tags.

Beta Features

  • Timeline Node: Arrange video/music clips into sequences for coherent multimedia works.
  • Result Node: Preview and download outputs; copy text results.
6

Section 06

Multi-project & Theme Support

Multi-project Management

Save multiple projects with independent canvas states; switch between projects or view all in the project list.

Theme Options

  • Dark: Dark background with bright green accents (good for long sessions).
  • Gold: Warm cream background with dark green accents (elegant retro).
  • Light: Minimalist cold tone (clean and simple).

Quick Start

  1. Install dependencies: npm install
  2. Start backend (port3100) and frontend (port5173): npm start

Wallet Configuration

Shares ~/.blockrun/ directory with Franklin core:

  • Base chain: Uses ~/.blockrun/wallet
  • Solana chain: Uses ~/.blockrun/solana-wallet
7

Section 07

Application Scenarios & Value

Franklin Canvas is ideal for:

  1. Freelancers: Pay-per-use avoids subscription waste, perfect for project-based work.
  2. Model Testers: Compare multiple models on the same canvas quickly.
  3. Complex Workflows: Orchestrate multi-step processes (e.g., text2img → img2video → add music).
  4. Crypto-native Creators: Use USDC without binding credit cards or bank accounts.

This platform provides lower trial costs and higher creative freedom for users, while opening new user groups for AI service providers.

8

Section 08

Conclusion: AI Creation Meets Web3 Payment

Franklin Canvas represents a new direction in combining AI media creation with Web3 payment. The x402-based pay-per-use model breaks traditional subscription constraints, offering flexible and cost-effective options.

For creators: Lower trial-and-error costs and more creative freedom. For providers: Access to users who prefer on-demand payment over subscriptions.

Overall, it's a promising tool bridging AI and Web3, catering to modern creators' evolving needs.