Meme Generator

Create meme videos by combining video or static images with audio. Runs entirely in Docker - no dependencies to install on your machine.

Features

• Video + Audio — Merge any video file with a separate audio track
• Image + Audio — Turn a static image (PNG, JPEG, BMP, TIFF, WebP) into a video with audio
• Tweet + Audio — Download a video from a Twitter/X post and use it as input
• Animated GIF + Audio — Use animated GIFs as video input, preserving their animation
• Meme captions — Add top and/or bottom text overlays in classic meme style
• Video/GIF looping — Automatically loops short videos or GIFs to fill the specified duration
• Audio looping — Loop short audio clips to match the video length
• Custom duration — Trim or extend output to an exact length in seconds
• YouTube audio download — Download audio from YouTube videos using yt-dlp
• Audio trimming — Extract a specific time range from downloaded audio (start/end points)
• MP4 output — H.264 video + AAC audio, optimized for social media with faststart
• Fully Dockerized — No local dependencies required beyond Docker

Example

angel_wojak_crying.mp4

Prerequisites

• Docker

Build

docker build -t meme-maker .

Usage

All commands use volume mounts to pass files in and get the result out. The output file is written inside the container to /output/ which you mount to a local directory so the file appears on your machine.

Video + Audio

Merge a video file with a separate audio track.

docker run --rm \
  -v /absolute/path/to/inputs:/input:ro \
  -v /absolute/path/to/output:/output \
  meme-maker \
  --input /input/clip.mp4 \
  --audio /input/song.mp3

The result will be at /absolute/path/to/output/meme.mp4.

Custom output filename

docker run --rm \
  -v /absolute/path/to/inputs:/input:ro \
  -v /absolute/path/to/output:/output \
  meme-maker \
  --input /input/clip.mp4 \
  --audio /input/song.mp3 \
  --output /output/my-meme.mp4

Specify duration

Trim or cap the output video to a specific length in seconds.

docker run --rm \
  -v /absolute/path/to/inputs:/input:ro \
  -v /absolute/path/to/output:/output \
  meme-maker \
  --input /input/clip.mp4 \
  --audio /input/song.mp3 \
  --duration 10

Loop short audio

If your audio clip is shorter than the video, loop it to fill the entire duration.

docker run --rm \
  -v /absolute/path/to/inputs:/input:ro \
  -v /absolute/path/to/output:/output \
  meme-maker \
  --input /input/clip.mp4 \
  --audio /input/song.mp3 \
  --loop-audio

Image + Audio

Create a video from a static image and an audio file. The image is displayed for the entire duration of the audio (or a custom duration).

docker run --rm \
  -v /absolute/path/to/inputs:/input:ro \
  -v /absolute/path/to/output:/output \
  meme-maker \
  --input /input/photo.png \
  --audio /input/voiceover.mp3

The image format can be PNG, JPEG, BMP, TIFF, or WebP. Duration defaults to the length of the audio file.

Image + Audio with custom duration

docker run --rm \
  -v /absolute/path/to/inputs:/input:ro \
  -v /absolute/path/to/output:/output \
  meme-maker \
  --input /input/photo.jpg \
  --audio /input/sound.mp3 \
  --duration 15

This creates a 15-second video of the static image with the first 15 seconds of the audio.

Tweet + Audio

Download the embedded video from a Twitter/X post and combine it with an audio track. Uses yt-dlp under the hood.

docker run --rm \
  -v /absolute/path/to/inputs:/input:ro \
  -v /absolute/path/to/output:/output \
  meme-maker \
  --tweet "https://x.com/user/status/1234567890" \
  --audio /input/song.mp3

The --tweet flag is mutually exclusive with --input — use one or the other. All other options (--duration, --loop-audio, --top-text, --bottom-text, etc.) work with --tweet too.

Tweet with original audio, trimmed, and captioned

Download a tweet video, trim to a specific time range, keep the original audio, and add a caption — all in one command:

docker run --rm \
  -v /absolute/path/to/output:/output \
  meme-maker \
  --tweet "https://x.com/user/status/1234567890" \
  --start 00:01:27 \
  --end 00:01:47 \
  --top-text "WHEN AGI TAKES OVER"

When --audio is omitted, the original video audio is preserved.

Meme captions

Add classic meme-style text (white with black outline) to the top and/or bottom of the video.

docker run --rm \
  -v /absolute/path/to/inputs:/input:ro \
  -v /absolute/path/to/output:/output \
  meme-maker \
  --input /input/clip.mp4 \
  --audio /input/song.mp3 \
  --top-text "WHEN THE CODE COMPILES" \
  --bottom-text "ON THE FIRST TRY"

You can use just --top-text, just --bottom-text, or both. The font size scales automatically with the video resolution. Special characters (colons, apostrophes, semicolons) are escaped automatically.

Full example (end to end)

Assuming your files are in ~/meme-files/ and you want the output in the current directory:

# Build once
docker build -t meme-maker .

# Video + audio, 30 second meme
docker run --rm \
  -v ~/meme-files:/input:ro \
  -v "$(pwd)/output":/output \
  meme-maker \
  --input /input/funny-clip.mp4 \
  --audio /input/bass-drop.mp3 \
  --duration 30 \
  --output /output/final-meme.mp4

# Image + audio, full audio length
docker run --rm \
  -v ~/meme-files:/input:ro \
  -v "$(pwd)/output":/output \
  meme-maker \
  --input /input/reaction-face.png \
  --audio /input/dramatic-sound.mp3 \
  --output /output/image-meme.mp4

Output files appear in ./output/ on your host machine.

Download Audio from YouTube

Extract audio from a YouTube video and save it as an MP3. This uses yt-dlp which is included in the Docker image.

docker run --rm \
  -v /absolute/path/to/output:/output \
  --entrypoint /app/download-audio.sh \
  meme-maker \
  --url "https://www.youtube.com/watch?v=VIDEO_ID"

The audio file will be saved to /absolute/path/to/output/audio.mp3.

Custom output filename

docker run --rm \
  -v /absolute/path/to/output:/output \
  --entrypoint /app/download-audio.sh \
  meme-maker \
  --url "https://www.youtube.com/watch?v=VIDEO_ID" \
  --output /output/my-song.mp3

Trim to a specific time range

Download audio and extract only a portion using --start and --end. Times can be in HH:MM:SS format or plain seconds.

docker run --rm \
  -v /absolute/path/to/output:/output \
  --entrypoint /app/download-audio.sh \
  meme-maker \
  --url "https://www.youtube.com/watch?v=VIDEO_ID" \
  --start 00:01:05 \
  --end 00:01:45

This downloads the full audio, then trims it to the 1:05–1:45 range (40 seconds) and saves the result.

You can also use just --start (trim from that point to the end) or just --end (trim from the beginning to that point):

# First 30 seconds only
docker run --rm \
  -v /absolute/path/to/output:/output \
  --entrypoint /app/download-audio.sh \
  meme-maker \
  --url "https://www.youtube.com/watch?v=VIDEO_ID" \
  --end 30

Download audio then create a meme (end to end)

# Download audio from YouTube
docker run --rm \
  -v "$(pwd)/inputs":/output \
  --entrypoint /app/download-audio.sh \
  meme-maker \
  --url "https://www.youtube.com/watch?v=VIDEO_ID" \
  --output /output/song.mp3

# Use the downloaded audio to create a meme
docker run --rm \
  -v "$(pwd)/inputs":/input:ro \
  -v "$(pwd)/output":/output \
  meme-maker \
  --input /input/photo.jpg \
  --audio /input/song.mp3 \
  --duration 15

Options

make-meme.sh (default entrypoint)

Flag	Required	Description
--input	Yes*	Path to video or image file (inside the container, e.g. /input/file.mp4)
--tweet	Yes*	URL of a Twitter/X post with an embedded video (alternative to --input)
--audio	No**	Path to audio file. Omit to keep original video audio
--output	No	Output path (default: /output/meme.mp4)
--top-text	No	Text to display at the top of the video (white with black outline)
--bottom-text	No	Text to display at the bottom of the video (white with black outline)
--start	No	Start time for input video trimming (e.g. 00:01:30 or 90)
--end	No	End time for input video trimming (e.g. 00:02:15 or 135)
--duration	No	Duration in seconds. Defaults to the video length or audio length (for images)
--loop-audio	No	Loop the audio track if it is shorter than the video

* Either --input or --tweet is required, but not both. ** Required when using a static image as input. Optional for video inputs (omit to keep original audio).

download-audio.sh (via --entrypoint)

Flag	Required	Description
--url	Yes	YouTube video URL
--output	No	Output path (default: /output/audio.mp3)
--start	No	Start time for trimming (e.g. 00:01:30 or 90)
--end	No	End time for trimming (e.g. 00:02:15 or 135)

Output format

All output is MP4 with:

• Video: H.264, yuv420p pixel format, CRF 23
• Audio: AAC at 192 kbps
• faststart flag enabled for quick playback on social media

License

This project is licensed under the MIT License. See the LICENSE file for details.