Dynamic Files

Flexible handling of compressed folders, sequence patterns, and multi-output commands.

Rendi supports FFmpeg commands that work with dynamic inputs and outputs - cases where the number of files isn’t known ahead of time or is determined by the command itself.

Supported Dynamic Patterns

Pattern	Type	Description
%0d / %03d	Input/Output	Sequence patterns for numbered files (e.g., frame_%03d.png)
tee	Output	Write to multiple outputs simultaneously
segment	Output	Split output into multiple segment files
image2	Input	Read image sequences with glob patterns
concat	Input	Concatenate multiple files from a list
HLS/DASH	Input/Output	Streaming manifests with multiple segments

To use these patterns with Rendi, you’ll typically need compressed folders for input or the OUTPUT_FOLDER flag for output.

---

Dynamic Inputs

Rendi supports using a compressed .zip folder as an FFmpeg input source. This enables commands that require multiple input files, such as image sequences (%0d patterns), HLS playlists, or concat demuxer lists.

Use the field input_compressed_folder to supply a URL pointing to a .zip file. Rendi automatically downloads and decompresses the archive before running your FFmpeg command.

Important: Command run time includes the time required to decompress the .zip file.

Using custom fonts

To use custom fonts or auxiliary files (e.g., subtitles, watermarks), bundle everything together in a .zip and use input_compressed_folder.

Important: When using input_compressed_folder, all input files must be inside the ZIP — do not use the input_files field alongside it.

For example, to add subtitles with a custom font to a video, create a ZIP containing:

• sample_popeye_meets_sinbad.mp4 - your video file
• sample_popeye_meets_sinbad.srt - your subtitle file
• Poppins-Regular.ttf - a font that supports your target script

curl --location 'https://api.rendi.dev/v1/run-ffmpeg-command' \
--header 'X-API-KEY: <api-key>' \
--header 'Content-Type: application/json; charset=utf-8' \
--data '{
  "input_compressed_folder": "https://storage.rendi.dev/sample/popeye_meets_sinbad_custom_font.zip",
  "output_files": {
    "out_1": "sample_popeye_meets_sinbad_subtitled.mp4"
  },
  "ffmpeg_command": "-i sample_popeye_meets_sinbad.mp4 -vf \"subtitles=sample_popeye_meets_sinbad.srt:fontsdir=.:force_style='\''FontName=Poppins,FontSize=24,PrimaryColour=&HFFFFFF,OutlineColour=&H4066B66B,Outline=1,BorderStyle=3'\''\" -c:v libx264 -c:a copy {{out_1}}"
}'

The FFmpeg command uses:

• -i sample_popeye_meets_sinbad.mp4 - Reads the video file from the decompressed ZIP
• subtitles=sample_popeye_meets_sinbad.srt - References the subtitle file from the same decompressed ZIP
• fontsdir=. - Tells FFmpeg to look for fonts in the current (decompressed) directory
• force_style='FontName=Poppins,...' - Applies the custom font and styling. Use the font’s FontName (found by opening the font file), not the filename

This approach works for any use case where FFmpeg needs to reference auxiliary files like fonts, text files, subtitle files, or watermark images — bundle everything into a single ZIP.

Tip: For text overlays with custom fonts, use drawtext=fontfile=MyFont.ttf:text='Hello':fontcolor=white:fontsize=60. See the FFmpeg Cheatsheet for more details.

Using HLS recording as input stored in a ZIP folder

This example demonstrates processing an HLS recording stored in a ZIP archive. The sample ZIP at playlist_sample.zip contains:

• playlist.m3u8 - the HLS playlist file
• output0.ts, output1.ts, output2.ts - the video segments

curl --location 'https://api.rendi.dev/v1/run-ffmpeg-command' \
--header 'X-API-KEY: <api-key>' \
--header 'Content-Type: application/json; charset=utf-8' \
--data '{
  "input_compressed_folder": "https://storage.rendi.dev/sample/playlist_sample.zip",
  "output_files": {
    "out_1": "recording.mp4"
  },
  "ffmpeg_command": "-i playlist.m3u8 -c:v libx264 -crf 23 -c:a aac -b:a 128k {{out_1}}",
  "vcpu_count": 8
}'

The FFmpeg command reads the HLS playlist, re-encodes the video to H.264 (CRF 23 for balanced quality/size) and audio to AAC, then outputs a single MP4 file.

Tip: For faster processing without re-encoding, use -c copy -bsf:a aac_adtstoasc instead.

Using indexed image frames to create a video

This example demonstrates creating a video from a sequence of numbered image frames stored in a ZIP archive. The ZIP should contain files like frame_001.png, frame_002.png, etc.

curl --location 'https://api.rendi.dev/v1/run-ffmpeg-command' \
--header 'X-API-KEY: <api-key>' \
--header 'Content-Type: application/json; charset=utf-8' \
--data '{
  "input_compressed_folder": "https://storage.rendi.dev/sample/big_buck_bunny_720p_5sec_30fps_frames.zip",
  "output_files": {
    "out_1": "video.mp4"
  },
  "ffmpeg_command": "-framerate 30 -i frame_%03d.png -c:v libx264 -pix_fmt yuv420p {{out_1}}",
  "vcpu_count": 8
}'

The FFmpeg command uses:

• -framerate 30 - Sets the input frame rate to 30 FPS
• -i frame_%03d.png - Reads sequentially numbered frames (frame_001.png, frame_002.png, …)
• -c:v libx264 - Encodes output as H.264 video
• -pix_fmt yuv420p - Ensures compatibility with most video players

Tip: Adjust the pattern to match your file naming (e.g., img_%04d.jpg for img_0001.jpg, img_0002.jpg, etc.).

Dynamic Outputs

Rendi also supports compressed output folders. To enable this, set the output_files field to the special flag:

OUTPUT_FOLDER

When this flag is used, Rendi creates a .zip archive containing only the output files produced by your FFmpeg command.

Important: Command runtime includes the time required to compress the output folder.

Extracting frames from an input video and storing them in a zip

This example extracts individual frames from a video and packages them into a ZIP archive. The output can be used as input for the “Using indexed image frames to create a video” example above.

curl --location 'https://api.rendi.dev/v1/run-ffmpeg-command' \
--header 'X-API-KEY: <api-key>' \
--header 'Content-Type: application/json; charset=utf-8' \
--data '{
  "input_files": {
    "in_1": "https://storage.rendi.dev/sample/big_buck_bunny_720p_5sec.mp4"
  },
  "output_files": "OUTPUT_FOLDER",
  "ffmpeg_command": "-i {{in_1}} -vf fps=30 frame_%03d.png"
}'

The FFmpeg command uses:

• -i {{in_1}} - Reads the input video file
• -vf fps=30 - Extracts frames at 30 frames per second
• frame_%03d.png - Outputs frames as frame_001.png, frame_002.png, etc.

The resulting ZIP contains all extracted frames, ready to be used with input_compressed_folder for video reconstruction or other processing.

Tip: Adjust the fps value to control frame extraction density. Use fps=1 for 1 frame per second, or omit -vf fps=30 entirely to extract every frame at the video’s native framerate.

Generating an HLS output folder and storing it as a ZIP

curl --location 'https://api.rendi.dev/v1/run-ffmpeg-command' \
--header 'X-API-KEY: <api-key>' \
--header 'Content-Type: application/json; charset=utf-8' \
--data '{
    "input_files": {
        "in_1": "https://storage.rendi.dev/sample/sample.avi"
    },
    "output_files": "OUTPUT_FOLDER",
    "ffmpeg_command": "-i {{in_1}} -t 25 -c:v h264 -c:a aac -b:v 2000k -b:a 128k -hls_time 5 -hls_list_size 0 -f hls playlist.m3u8"
}'

Response Format for Compressed Output Folders

In addition to the usual fields, the response includes extra metadata describing the generated .zip archive:

{
  "command_id": "41938bd7-3550-46d8-87c3-9d0b3200e7fb",
  "status": "SUCCESS",
  "command_type": "FFMPEG_COMMAND",
  "total_processing_seconds": 31.964526,
  "ffmpeg_command_run_seconds": 3.218038558959961,
  "vcpu_count": 8,
  "output_files": {
    "OUTPUT_FOLDER": {
      "file_id": "8d9ebcfb-f560-4858-af5c-bd5bf426d874",
      "storage_url": "https://storage.rendi.dev/files/.../41938bd7-3550-46d8-87c3-9d0b3200e7fb.zip",
      "status": "STORED",
      "rendi_store_type": "OUTPUT",
      "is_deleted": false,
      "size_mbytes": 6.900238990783691,
      "file_type": "zip",
      "file_format": "zip",
      "file_count": 6,
      "size_compressed_mbytes": 5.120238990783691
    }
  }
}

Field Descriptions

• file_count - Number of files contained inside the output .zip folder.
• size_mbytes - Total size (MB) of all files before compression.
• size_compressed_mbytes - Final compressed .zip size stored in Rendi.
• file_format - Always zip.
• storage_url - The download URL of the compressed output folder.

The compressed output file name is: <command_id>.zip

---

Processing & Storage Quota

1. Processing quota for inputs is calculated based on the size of the files after decompression.
2. Processing quota for OUTPUT_FOLDER uses the size of files before compression.
3. Storage usage is based on the size of the compressed .zip file.