screencastgen

Web Config

Application settings from environment variables.

Source: web/backend/config.py

Class: `Settings`

All environment variables use the P2A_ prefix.

Variable	Type	Default	Description
`P2A_DATABASE_URL`	`str`	—	PostgreSQL async connection string
`P2A_SYNC_DATABASE_URL`	`str`	—	PostgreSQL sync connection for Celery
`P2A_REDIS_URL`	`str`	—	Redis URL for broker/results
`P2A_UPLOAD_DIR`	`str`	`"./uploads"`	Upload file storage path (local backend)
`P2A_OUTPUT_DIR`	`str`	`"./outputs"`	Job output and pipeline working directory
`P2A_MAX_UPLOAD_SIZE_MB`	`int`	`200`	Maximum upload file size
`P2A_TTS_SERVER_URL`	`str`	`http://localhost:8100`	GPU inference server URL
`P2A_TTS_CONCURRENCY`	`int`	`8`	Number of synthesis requests the worker submits in parallel to the TTS server
`P2A_ALLOWED_ORIGINS`	`List[str]`	`["http://localhost:5173"]`	CORS allowed origins
`P2A_STORAGE_BACKEND`	`str`	`"local"`	Storage backend: `local`, `gcs`, or `s3`
`P2A_STORAGE_BUCKET`	`str`	`""`	Bucket name for GCS/S3 backends
`P2A_STORAGE_PREFIX`	`str`	`""`	Optional key prefix inside the bucket
`P2A_STORAGE_REGION`	`str`	`""`	AWS region for S3
`P2A_STORAGE_LOCAL_CACHE_DIR`	`str`	`"/tmp/screencastgen_cache"`	Local cache for remote storage downloads

Notes

TTS_CONCURRENCY works with the inference server’s batcher: the web worker sends multiple chunk requests concurrently, and the GPU server coalesces compatible requests into a single batched forward pass.

Usage

from web.backend.config import settings

server_url = settings.TTS_SERVER_URL
tts_concurrency = settings.TTS_CONCURRENCY

Singleton instance settings is created at import time.

Dependencies

Web Config
├── pydantic-settings
└──▶ consumed by FastAPI App
     ├──▶ Web Database
     ├──▶ Uploads Router
     ├──▶ Events Router
     ├──▶ Pipeline Tasks
     └──▶ Storage Service

screencastgen

Web Config

Class: Settings

Notes

Usage

Dependencies

See Also

Class: `Settings`