iTube is a YouTube clone-ish platform built as a set of small services: a Flutter client app, a FastAPI backend, an SQS-to-ECS consumer, and a transcoder that turns uploaded videos into HLS/DASH outputs.

This repository is an umbrella/index repo. It does not contain source code. It documents how the system fits together and links to the component repositories.

flowchart TB
  client["Client (Flutter/Web)"]
  api["iTube-api (FastAPI)"]
  s3raw["S3 (raw uploads)"]
  sqs["SQS queue"]
  consumer["iTube-consumer"]
  transcoder["iTube-transcoder"]
  s3processed["S3 (processed outputs)"]
  postgres["PostgreSQL"]
  redis["Redis"]

  client -->|"HTTP: auth/metadata/listing"| api
  api --> postgres
  api --> redis
  api -->|"Presigned S3 upload URLs"| client
  client -->|"Upload video/thumbnail"| s3raw
  s3raw -->|"S3 event"| sqs
  sqs --> consumer
  consumer -->|"ECS RunTask (Fargate)"| transcoder
  transcoder -->|"Upload HLS/DASH"| s3processed
  transcoder -->|"Status update (IAM)"| api

1. The client authenticates with the API (AWS Cognito) and requests presigned S3 upload URLs.
2. The client uploads raw video and thumbnail to S3.
3. S3 emits an object-created event to SQS.
4. iTube-consumer polls the SQS queue and starts an ECS Fargate task for iTube-transcoder, injecting the S3 bucket/key.
5. iTube-transcoder downloads the raw video, transcodes to HLS/DASH, and uploads processed outputs to the processed S3 bucket.
6. The transcoder calls back into the API to update processing status and metadata.
7. Clients query the API for video listings and playback metadata.

• AWS Cognito for user authentication.
• S3 buckets for raw uploads, processed outputs, and thumbnails.
• SQS queue for upload event fanout.
• ECS Fargate for running transcoder tasks.
• PostgreSQL for persistent metadata.
• Redis for caching.

Each component repo has its own README with local development and deployment details. Start with the component you want to run and follow its setup instructions.