FFmpeg Web Video Compressor

Web application for analyzing and compressing video files in the browser using Flask and FFmpeg.

The app is currently configured and tested primarily for this workflow:

• Hosted on a PC -> Accessed from Safari/browser on iOS over the local network

Requirements

Python

• Python 3.10+
• Packages in requirements.txt

Install Python dependencies:

pip install -r requirements.txt

Non-Python System Dependencies

These must be installed on the host machine and available in PATH:

• ffmpeg (video conversion)
• ffprobe (video metadata analysis)
• exiftool (metadata copy/restore)

Linux (Debian/Ubuntu):

sudo apt update
sudo apt install -y ffmpeg libimage-exiftool-perl

Verify binaries:

ffmpeg -version
ffprobe -version
exiftool -ver

Notes:

• ffprobe is usually installed with ffmpeg.
• exiftool is required by the metadata restore flow.
• The ffmpeg and ffprobe version should not matter much, but exiftool requires version >=13.13, as it added the ability to properly write lens info for iOS metadata

Run

Start the app:

python3 app.py

Optional EXIF dump logging:

python3 app.py --enable-exif-logging

Default bind address:

• 0.0.0.0:5000

Access it at:

• https://<host-ip>:5000

RAM Storage On Linux And Windows

The app writes uploads and converted outputs to a temp directory selected by the backend. Startup uses modules/configureos.py to detect the OS and auto-apply runtime env setup.

By default, RAM-backed storage is required (FFMPEG_WEB_REQUIRE_RAM_STORAGE=1). If RAM-backed storage cannot be found, the app exits at startup.

Priority order:

• FFMPEG_WEB_RAMDISK_DIR (if set)
• /dev/shm (Linux)
• /run/shm (Linux)
• OS default temp folder

Note:

• The OS default temp folder is only used if RAM requirement is disabled (FFMPEG_WEB_REQUIRE_RAM_STORAGE=0).
• The app creates a per-run temp workspace directory and automatically removes it on process exit.
• Flask upload spool files are also forced into the same RAM-backed workspace.
• The app sets TMPDIR/TMP/TEMP to the active RAM workspace at startup.

Linux

No extra setup is needed if /dev/shm is available. The app auto-sets FFMPEG_WEB_RAMDISK_DIR to /dev/shm (or /run/shm fallback) when unset.

Windows (RAM-backed mode)

Windows does not provide /dev/shm, so the app can auto-create a RAM disk if ImDisk is installed.

1. Install ImDisk Toolkit (must expose imdisk command in PATH).
2. Run the app normally.

The app will auto-create and mount a RAM disk on R: (default), create R:\ffmpeg-temp, and use it as FFMPEG_WEB_RAMDISK_DIR. If the app created the RAM disk, it will detach it on process exit.

Optional Windows tuning:

setx FFMPEG_WEB_WINDOWS_RAMDISK_SIZE "8G"
setx FFMPEG_WEB_WINDOWS_RAMDISK_DRIVE "R:"
setx FFMPEG_WEB_WINDOWS_RAMDISK_FOLDER "ffmpeg-temp"

You can still override directly:

setx FFMPEG_WEB_RAMDISK_DIR "R:\\ffmpeg-temp"

If FFMPEG_WEB_RAMDISK_DIR is not set (or points to a missing folder), startup fails in default RAM-only mode.

Optional fallback mode (allows disk-backed temp):

setx FFMPEG_WEB_REQUIRE_RAM_STORAGE "0"

Minimal Cross-Platform Setup

Linux:

1. Install dependencies from this README.
2. Ensure /dev/shm exists (default on most distros).
3. Run python3 app.py.

Windows:

1. Install dependencies (ffmpeg, ffprobe, exiftool).
2. Install ImDisk Toolkit (for automatic RAM-disk creation).
3. Run python app.py.

API Endpoints

• GET / main page
• POST /analyze upload and analyze a video
• POST /convert convert uploaded video
• GET /download/<filename> download converted video
• GET /cache/status cache status
• GET /hardware/status hardware encoder status
• GET /progress/<file_id> conversion progress snapshot
• GET /progress/stream/<file_id> conversion progress SSE stream
• POST /cleanup delete cached files