Cloud FFmpeg API
Run native FFmpeg commands through an API or Agent Skill without tying up your laptop or servers.
# one-time setup, or: ffhub config <key> $ export FFHUB_API_KEY=sk_… $ npx ffhub -i https://storage.ffhub.io/sample.mp4 -c:v libx264 output.mp4 Command: ffmpeg -i … -c:v libx264 output.mp4 Task created: 01abcd… Processing...
Progress: 100% [running] Done! Execution time: 12.4s Downloading output.mp4 (3.1 MB, Ctrl+C to cancel)
swipe code horizontally →
How it works
From FFmpeg command to output URL in four steps
Create an API key
Sign up, generate a key, start with trial credits.
Submit an FFmpeg task
POST a native FFmpeg command with a public input URL.
Track task status
Tasks run async: pending → running → succeeded. Poll the task or use a webhook.
Download the output
Completed tasks return the URL of the processed file.
Native FFmpeg commands
Send the same flags you already run locally. No proprietary DSL to learn.
Async task processing
Long jobs never block your request. Every task has an id and a status.
File upload or input URL
Upload directly, or point FFHub at a URL your storage already serves.
Status and output URL
One task object carries progress, errors and the finished file location.
No FFmpeg servers to run
No build flags, no codec installs, no worker fleet, no autoscaling.
Reference
Agent Skills
Let your AI agent run FFmpeg in the cloud
Install the open-source FFHub FFmpeg Skill and describe media tasks in natural language. Your agent generates native FFmpeg commands, uploads inputs, and calls the FFHub cloud API—without installing or running FFmpeg locally.
Works with Claude Code, Codex, Cursor, OpenCode, and 40+ tools that support Agent Skills.
Describe the result you want
Say “compress this MOV for the web” without looking up FFmpeg flags first.
The Skill handles the workflow
It detects local files, uploads inputs, builds the command, and calls the FFHub cloud API.
Get the finished file back
It tracks task progress and returns the output URL with file details.
Using a general assistant such as ChatGPT or Claude.ai? Point it to llms.txt for accurate FFHub integration context.
Familiar FFmpeg flags for common media tasks
Video compression
-c:v libx264 -crf 26
Format conversion
mp4 · webm · mov · gif
Audio extraction
-vn -c:a libmp3lame
Trim and cut
-ss 00:00:10 -t 15
Resize and scale
-vf scale=1280:-2
Thumbnails
-frames:v 1 cover.jpg
Validate online, then connect the API
Check the result in a browser tool, then send the same command through the API for automation.
Pay only for actual processing time
Start with trial credits. After that, credits are charged by actual FFmpeg execution time — no seats or monthly minimum.
See pricing →Run your first cloud task in five minutes
Follow the request flow, or generate a key and send a command.