root
2f5291a0f9
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.
img-infinite-canvas API
go-zero backend for the Moteva-style infinite canvas MVP.
Features
- Dashboard data for the Moteva home page
- Project creation from prompt
- Project title rename
- Full canvas project reads
- Moteva-style generation mock for image/design workflows
- Canvas layout save
- Chat history replay
- DDD layout with domain, application, infrastructure, and generated handler layers
- sky-valley/pi Go agent adapter with local desktop agent disabled
- Optional PostgreSQL persistence through pgx/sqlc
- Optional Redis cache, default memory cache
- Asynq job queue for generation, OCR, layer merge, and asset cleanup workers
- Image asset upload URLs for memory, MinIO, AWS S3, Cloudflare R2, and Aliyun OSS
Configuration
Name: img_infinite_canvas-api
Host: 0.0.0.0
Port: 8888
Timeout: 1200000
Storage:
Driver: postgres
DataSource: postgres://canvas:canvas@localhost:15432/canvas?sslmode=disable
Cache:
Driver: redis
Addr: localhost:16379
Password: ""
DB: 0
TTLSeconds: 300
JobQueue:
Driver: asynq
Addr: localhost:16379
Password: ""
DB: 0
Queue: default
Concurrency: 10
MaxRetry: 3
TimeoutSeconds: 900
ShutdownTimeoutSeconds: 30
Agent:
Driver: pi
Text:
Driver: vision
Model: gpt-5.4-mini
TimeoutSeconds: 240
CacheTTLSeconds: 600
Planner:
Driver: vision
Model: gpt-5.5
TimeoutSeconds: 240
Image:
BaseURL: http://154.21.87.3:8080
APIKeyEnv: GPT_IMAGE_API_KEY
InputImageTransport: file
TimeoutSeconds: 600
ObjectStorage:
Driver: minio
Bucket: canvas-assets
Endpoint: localhost:19000
Region: us-east-1
AccessKey: canvas
SecretKey: canvas-secret
PublicBase: http://localhost:19000/canvas-assets
UseSSL: false
ExpiresIn: 900
Agent.Image.InputImageTransport controls how input/reference images are sent to the image model for edit/image-to-image calls: file uploads image bytes with multipart form data, while url sends images: [{"image_url": "..."}] for gateways that can fetch public URLs directly.
Start local Docker dependencies for debugging:
docker compose -f ../deploy/docker-compose.yml up -d postgres redis minio
Use PostgreSQL, Redis, and MinIO inside Docker Compose. The compose file runs the API and Asynq worker as separate processes:
Storage:
Driver: postgres
DataSource: postgres://canvas:canvas@postgres:5432/canvas?sslmode=disable
Cache:
Driver: redis
Addr: redis:6379
TTLSeconds: 300
JobQueue:
Driver: asynq
Addr: redis:6379
Queue: default
TimeoutSeconds: 900
ObjectStorage:
Driver: minio
Bucket: canvas-assets
Endpoint: minio:9000
Region: us-east-1
AccessKey: canvas
SecretKey: canvas-secret
PublicBase: http://localhost:19000/canvas-assets
UseSSL: false
Running
go run . -f etc/config.yaml -mode all
go run . -f etc/config.yaml -mode api
go run . -f etc/config.yaml -mode worker
Testing
go test ./...
go build ./...
Example
curl http://localhost:8888/api/dashboard
curl -X POST http://localhost:8888/api/projects \
-H "Content-Type: application/json" \
-d '{"prompt":"为 still day 品牌服装独立站设计官网视觉方向"}'
curl -X PATCH http://localhost:8888/api/projects/<project-id>/title \
-H "Content-Type: application/json" \
-d '{"title":"夏季活动主视觉"}'
curl -X POST http://localhost:8888/api/assets/upload-url \
-H "Content-Type: application/json" \
-d '{"fileName":"ref.png","contentType":"image/png"}'
Detailed endpoint docs are in API.md.