Run FFmpeg in n8n and on n8n Cloud
n8n Cloud has no Execute Command node, so a workflow there can't start FFmpeg. Install the verified Rendobar node, give it any FFmpeg command, and the finished file comes back to your workflow. It works the same on n8n Cloud and on your own instance.
- Guide
- FFmpeg API
- Updated
- November 2026
{ "name": "Run FFmpeg on Rendobar", "type": "@rendobar/n8n-nodes-rendobar.rendobar", "typeVersion": 1, "parameters": { "resource": "job", "operation": "create", "jobType": { "__rl": true, "mode": "id", "value": "ffmpeg" }, "inputsMode": "json", "inputs": "={{ JSON.stringify({ source: $json.videoUrl }) }}", "paramsMode": "fields", "params": { "mappingMode": "defineBelow", "value": { "command": "-i source -vf scale=-2:720 -c:v libx264 -crf 28 output.mp4" } }, "options": { "callbackUrl": "={{ $execution.resumeUrl }}" }, "output": "simplified" }}Versus
n8n Cloud vs Rendobar
| Topic | n8n Cloud | |
|---|---|---|
| FFmpeg | No way to run it The Execute Command node isn't available on Cloud | Any command The same flags you run locally |
| CPU | 10 to 80 millicores Burstable, set by plan | 64 cores |
| Memory | 640 MiB On Trial, Starter and Pro-1. 1280 MiB on Pro-2, 4096 MiB on Enterprise. | 336 GB |
| Disk | 100 GB For the whole instance | 512 GB Per job |
| Video in | Counts against memory Binary data size raises memory use | A URL Inputs up to 10 GB on Pro and Studio |
n8n figures from the Execute Command node, Cloud data management and memory errors pages, checked 16 November 2026.
n8n Cloud runs no FFmpeg, on a CPU or a GPU. An NVENC command runs on one of these, on Pro and Studio.
NVIDIA L4
- Memory
- 24 GB
- NVENC engines
- 2
NVIDIA L40S
- Memory
- 48 GB
- NVENC engines
- 3
NVIDIA RTX PRO 6000
- Memory
- 96 GB
- NVENC engines
- 4
How
How it works
Install the node
In any workflow, press + to open the nodes panel and search for Rendobar. It's verified by n8n, so it installs on n8n Cloud straight from that panel. Then add a Rendobar credential with an API key from the API Keys page.
n8n nodes panel - Search
- Rendobar
- Listed under
- More from the community
- Badge
- Verified
- Package
- @rendobar/n8n-nodes-rendobar
- Credential
- rb_••••••••
Send the FFmpeg command
That's the node at the top of this page. Inputs (JSON) gives the video's URL the name source, and the command uses that name where a file path would go. Any command from the recipes works the same way. The node returns the job at once.
What Job > Create returns - id
- job_8f2c41d07a6e4b19
- type
- ffmpeg
- status
- waiting
Park on a Wait node
The Create node's Callback URL is the Wait node's resume address, so Rendobar calls it when the job ends. Set the Wait node's HTTP Method to POST, since its default GET never resumes. A wait of 65 seconds or more is parked in n8n's database, so no worker is held while FFmpeg runs. Limit Wait Time gives a parked execution a way out, set past your longest job.
Wait node {"name": "Wait for the callback","type": "n8n-nodes-base.wait","typeVersion": 1.1,"parameters": {"resume": "webhook","httpMethod": "POST","limitWaitTime": true,"limitType": "afterTimeInterval","resumeAmount": 10,"resumeUnit": "hours"}}Get the file
n8n hands the resume call over as headers, params, query and body, so the job ID is at $json.body.data.jobId. Get fetches the job again over your own credential and puts the output on the item as binary data, ready for a Google Drive, Slack, S3 or email node.
Rendobar node, Job > Get {"name": "Fetch job and download file","type": "@rendobar/n8n-nodes-rendobar.rendobar","typeVersion": 1,"parameters": {"resource": "job","operation": "get","jobId": { "__rl": true, "mode": "id", "value": "={{ $json.body.data.jobId }}" },"downloadOutput": true,"outputBinaryProperty": "data","output": "simplified"}}
Templates
5 n8n video workflows to import
Every one runs on n8n Cloud and parks on a Wait node until the job finishes. Import one with Workflows > Import from File, add your Rendobar credential, and connect the other apps it names. They're all in the templates folder on GitHub, and the node is listed on n8n's Rendobar integration page. Compression, captions and image workflows are on the Rendobar n8n integration page.
FAQ
Frequently asked questions
Can I run FFmpeg on n8n Cloud?
Yes, with the Rendobar node. The Execute Command node isn't available on n8n Cloud, so a workflow there has nothing to start a binary with. The Rendobar node sends the command to Rendobar's FFmpeg API instead: the job runs any FFmpeg command and the finished file comes back to the workflow.
Why does my n8n workflow say ffmpeg: not found?
The n8n process has no FFmpeg binary to run. n8n Cloud has no Execute Command node at all, and from n8n 2.0 the node is off by default on every instance. The Rendobar node needs neither. Your command runs on Rendobar with the full FFmpeg build, and the output arrives on the item as a file.
Why can't I find the Rendobar node on n8n Cloud?
Search for Rendobar in the nodes panel. It's listed under More from the community with a verified badge, and the instance owner or an admin installs it for everyone. If it doesn't show, the instance owner turns on Verified Community Nodes in the Cloud admin panel, and the first time the instance may need a restart. The node's page on n8n is n8n.io/integrations/rendobar.
How long can an FFmpeg job run from n8n?
Up to 1 hour on Free and 9 hours on Pro and Studio. The workflow doesn't stay open for it. The job carries the Wait node's resume address as its callback, n8n parks the execution in its database, and Rendobar's call picks it back up when the job ends. Set Limit Wait Time past your longest job so a callback that never lands can't park the execution for good. Parking an n8n job on a Wait node walks through its two settings and a measured compression run.
How do I write the FFmpeg command in the node?
Write it the way you'd type it in a terminal, with the input's name where the file path goes. Inputs (JSON) maps names to URLs, for example source to your video, and the command reads -i source ... output.mp4. The recipes are written for exactly this, so trims, GIFs, watermarks and audio extraction are a copy away.
How do I process a video from Google Drive or another n8n node?
Put File > Upload after the node that has the file. It sends the binary to Rendobar and returns a URL for the Create node's inputs. The upload goes in 100 MB parts, so a large video never has to fit in n8n's memory, and uploaded files delete themselves after 24 hours. The Cut long videos into captioned vertical shorts template shows the whole chain.
How do I see why an FFmpeg command failed?
Start a second workflow from the Rendobar Trigger on Job Failed and follow it with Job > Get Logs. The whole runner log arrives on one item, FFmpeg's error text included, ready to post to Slack or email. The Run FFmpeg commands on n8n Cloud template already has this branch.
Can an n8n AI Agent run FFmpeg?
Yes. Connect the Rendobar node to the agent's Tool port and the agent picks the job type and writes the parameters. The job types come from your account when the agent looks, so a type added to Rendobar later is there with no node update.
What does it cost?
Jobs are billed per second of compute. A VP9 transcode we measured used 2.2 cores for 6.1 seconds and cost $0.0071. Every account starts with $5 of free credit and no card, and Rendobar bills nothing while you're not running jobs. Plans are on the pricing page.
Your command. Our infrastructure.
Bring your own commands. Run them on scalable CPU and GPU infrastructure, and pay only for compute time.
