root 2f5291a0f9 feat(image): pass reference images to GPT image model for edits
Server: forward canvas/reference images to the image model so extract,
continue, and edit requests stay faithful to the source product.
- Add ImageRequestOptions.Images and route requests with inputs to the
  /v1/images/edits endpoint, via multipart upload (default) or JSON
  image_url payload, selectable through Agent.Image.InputImageTransport.
- Extract reference URLs from prompt directives, inline @image tokens,
  and chat message image/mask contents; fold them into the idempotency
  key and prepend a reference-fidelity instruction to the prompt.

Frontend:
- Shrink CanvasToast to a compact top pill.
- Paste an internal transparent clipboard node instead of re-uploading
  when it matches the clipboard image.
- Keep agent-project polling alive on stream error while streaming.
2026-07-08 01:19:20 +08:00

img-infinite-canvas

Moteva 风格的 AI 设计工作台复刻 MVP:主页 Prompt 创建项目,项目页提供无限画布、节点拖拽、缩放平移、右侧设计对话、历史回放、图片资产上传、项目保存与再生成。

Stack

  • Backend: go-zero, DDD layering, sqlc, pgx, PostgreSQL optional persistence
  • Cache: memory or Redis, default memory
  • Agent: sky-valley/pi Go agent runtime adapter, local desktop agent disabled
  • Image storage: memory, MinIO, AWS S3, Cloudflare R2, Aliyun OSS
  • Frontend: Next.js App Router, pnpm, Vite artifact build, TypeScript, gzip, i18n
  • Deploy: local binary, Docker Compose, k3s manifests

Run Locally

Backend:

cd server
go mod tidy
go run . -f etc/config.yaml

Frontend:

cd frontend
pnpm install
pnpm dev

Open http://localhost:5173. Next.js rewrites /api to http://localhost:8888 in development. pnpm build runs next build and vite build, producing gzip assets in frontend/dist.

Cache Config

Default in-memory cache:

Cache:
  Driver: memory
  TTLSeconds: 300

Redis cache:

Cache:
  Driver: redis
  Addr: localhost:6379
  Password: ""
  DB: 0
  TTLSeconds: 300

Storage Config

Default memory storage is useful for local UI work:

Storage:
  Driver: memory

Image Object Storage

Default memory mode is useful for local prototyping. Production can use MinIO, AWS S3, Cloudflare R2, or Aliyun OSS:

ObjectStorage:
  Driver: minio # memory|minio|s3|aws|r2|aliyun
  Bucket: canvas-assets
  Endpoint: localhost:9000
  Region: us-east-1
  AccessKey: canvas
  SecretKey: canvas-secret
  PublicBase: http://localhost:9000/canvas-assets
  UseSSL: false
  ExpiresIn: 900

PostgreSQL storage:

Storage:
  Driver: postgres
  DataSource: postgres://canvas:canvas@localhost:5432/canvas?sslmode=disable

Docker Compose

cd deploy
docker compose up --build

Open http://localhost:3000.

k3s

Build and load/push these images first:

docker build -t img-infinite-canvas-api:latest server
docker build -t img-infinite-canvas-web:latest frontend
kubectl apply -f deploy/k3s/img-infinite-canvas.yaml

The sample ingress host is canvas.local.

API

See server/API.md.

S
Description
No description provided
Readme 161 MiB
Languages
TypeScript 44.5%
Go 43%
CSS 11.3%
HTML 1.1%