Feat/durable image processing architecture

README.md

@@ -52,8 +52,8 @@ content. Each placement has an authored position on its collection or folder
 canvas. Folders are not assets.
 
 The application consists of a React/Vite client, a Bun/Hono API backed by
-Postgres and Drizzle, plus an AWS Lambda image pipeline that processes S3
-uploads asynchronously through SQS. Collection and folder badges show descendant asset counts:
+Postgres and Drizzle, plus AWS Lambda image workers that process S3 uploads
+asynchronously through independent SQS queues. Collection and folder badges show descendant asset counts:
 images and notes count, folders do not. The client uses XYFlow for collection
 canvas rendering and interaction while Aska's API remains the source of truth
 for node identity, hierarchy, and position.

SST_DEPLOYMENT.md

@@ -6,7 +6,7 @@ One stage creates one isolated AWS copy:
 ```text
 stage dev
   API Gateway -> Hono Lambda
-  private S3 assets bucket -> SQS -> image-processing Lambda
+  private S3 assets bucket -> two SQS queues -> variants and palette Lambdas
   private S3 client bucket -> CloudFront -> React/Vite client
   dead-letter queue, IAM permissions, and stage-specific SST secrets
 ```
@@ -25,10 +25,10 @@ another API, bucket, queues, and Lambdas.
 | Direct package commands       | Your laptop                    | No AWS event chain                   | Unit tests and isolated debugging only                   |
 
 Both SST modes are real end-to-end AWS flows. With live development, an image
-uploaded from the browser goes to the real S3 bucket, creates a real SQS
-message, and invokes the image pipeline running locally. With a normal deploy,
-the same pipeline runs in AWS instead. Both test the actual permissions, event
-shape, queue flow, and callback path.
+uploaded from the browser goes to the real S3 bucket, creates one message in
+each image-processing queue, and invokes the variants and palette workers. With
+a normal deploy, those same workers run in AWS instead. Both test the actual
+permissions, event shape, queue flow, and callback path.
 
 ## One-time `dev` setup
 
@@ -180,7 +180,8 @@ this.
 3. Upload an image in the browser. It follows this real path:
 
    ```text
-   browser -> API -> S3 ingest/ -> SQS -> local image Lambda -> API callback
+   browser -> API -> S3 ingest/ -> variants SQS + palette SQS
+           -> local worker callbacks -> API
    ```
 
 When you stop `sst dev`, its AWS Lambda proxies can no longer reach your