Unrecognized node type: n8n-nodes-base.executeCommand means n8n 2.0 no longer loads Execute Command. Re-enable it with NODES_EXCLUDE, or have an HTTP Request node call ReelSDK at $0.02 per compute minute.
Unrecognized node type: n8n-nodes-base.executeCommand
We reproduced it on n8n 2.46 with a workflow that runs ffmpeg. An HTTP Request node calling ReelSDK finished the same 10-second encode in about 6 seconds, and one environment variable brings the old node back.
ErrorsUpdated
Error executing workflow. See log messages for details.
Execution error:
====================================
Unrecognized node type: n8n-nodes-base.executeCommand
UnrecognizedNodeTypeError: Unrecognized node type: n8n-nodes-base.executeCommand
at LazyPackageDirectoryLoader.getNode (.../n8n-core/src/nodes-loader/directory-loader.ts:258:9)
at LoadNodesAndCredentials.getNode (.../n8n/src/load-nodes-and-credentials.ts:725:29)
at NodeTypes.getByNameAndVersion (.../n8n/src/node-types.ts:111:45)
at new Workflow (.../n8n-workflow/src/workflow.ts:94:30)Source: n8n v2.0 breaking changes, checked November 2026.
Where it shows up
A workflow with an Execute Command node, on n8n 2.0 or later
You built a workflow that runs ffmpeg through an Execute Command node, then upgraded to n8n 2.0 or later, or imported the workflow into a new instance. Our test: n8n 2.46 from npm on Node.js 24, a Manual Trigger followed by Execute Command running ffmpeg -version, run with n8n execute.
With default settings it failed before any node ran: Unrecognized node type: n8n-nodes-base.executeCommand, with exit code 1. The stack ends in new Workflow, so n8n could not even build the workflow, let alone start the command.
Why it happens
n8n 2.0 no longer loads the node
n8n 2.0 disabled the Execute Command and Local File Trigger nodes by default, because they let workflow authors run any command and read the file system. They are switched off through the NODES_EXCLUDE variable, which now defaults to ["n8n-nodes-base.executeCommand", "n8n-nodes-base.localFileTrigger"].
An excluded node is never loaded, so to n8n its type does not exist, and every workflow that uses it fails with this error. On n8n Cloud there is no switch to flip: the docs say Execute Command isn't available there at all.
Causes and fixes
What we ran on n8n 2.46
| Setup | Result |
|---|---|
| Default settings, Execute Command | Unrecognized node type: n8n-nodes-base.executeCommand, exit code 1 |
| NODES_EXCLUDE without executeCommand, no ffmpeg | /bin/sh: 1: ffmpeg: not found |
| NODES_EXCLUDE without executeCommand, ffmpeg installed | Encoded to 1280x720 in about 7 s |
| HTTP Request node to POST /v1/ffmpeg | Same video, node done in about 6 s |
| n8n Cloud | Not tested: the docs say Execute Command isn't available |
Source: n8n v2.0 breaking changes, checked November 2026.
The local fix
Self-hosted: take the node off NODES_EXCLUDE, then install ffmpeg
Set NODES_EXCLUDE without the Execute Command node and restart n8n. NODES_EXCLUDE="[]" enables every node; we kept Local File Trigger blocked with NODES_EXCLUDE='["n8n-nodes-base.localFileTrigger"]' and the same workflow ran ffmpeg -version with exitCode 0.
The node then needs ffmpeg where n8n runs. Without it on PATH, our run failed with /bin/sh: 1: ffmpeg: not found, a NodeOperationError. With Debian's ffmpeg installed, the node scaled the 1080p demo clip down to 720p in about 7 seconds. The official Docker image has no apk, so the docs restore it from Alpine; the Dockerfile below follows their curl example, and we did not build it.
# n8n 2.46.6: load Execute Command, keep Local File Trigger off
export NODES_EXCLUDE='["n8n-nodes-base.localFileTrigger"]'
n8n startFROM n8nio/n8n
USER root
# Restore the apk package manager (removed from the base image)
COPY --from=alpine:3.22 /sbin/apk /sbin/apk
COPY --from=alpine:3.22 /lib/apk /lib/apk
COPY --from=alpine:3.22 /usr/lib/libapk* /usr/lib/
RUN apk add --no-cache ffmpeg
USER nodeSkip the binary
Or call ffmpeg from an HTTP Request node
No shell needed: an HTTP Request node can post the ffmpeg arguments to POST /v1/ffmpeg, which runs ffmpeg 8.1 on a fresh machine per job. We sent the request below from an HTTP Request node on n8n 2.46 with default settings, and the node finished in about 6 seconds with a 720p out.mp4. This is also the route on n8n Cloud, which we did not test.
Leave the node's Timeout option unset: it then waits 5 minutes, while adding the option starts it at 10 seconds, shorter than the 60 seconds the ?wait=60 request can hold. Compute is $0.02 per minute, billed per second, so this job cost under a fifth of a cent. For the full workflow, see our guide to ffmpeg in n8n.
curl "https://api.reelsdk.com/v1/ffmpeg?wait=60" \
-H "Authorization: Bearer $REELSDK_KEY" \
--json '{
"inputs": {
"in.mp4": "https://api.reelsdk.com/f/lib/demo/landscape.mp4"
},
"args": [
"-i", "in.mp4", "-vf", "scale=-2:720", "-c:v", "libx264", "-crf", "28",
"-preset", "veryfast", "-c:a", "aac", "-movflags", "+faststart",
"out.mp4"
],
"outputs": ["out.mp4"]
}'Which fix to pick
Re-enabling the node is quick on a server you control, but it gives every workflow author a shell on that machine, and each encode then uses its CPU. If others can edit workflows, keep the node off and use the HTTP Request route; on n8n Cloud that is the only route.
If you do re-enable it, list only what you need instead of [], so Local File Trigger stays off, and pin your n8n version so an upgrade does not change the default under you.
FAQ
Questions, answered
Why did my Execute Command node disappear after upgrading n8n?
n8n 2.0 added it to NODES_EXCLUDE by default, so it is not loaded and workflows that use it fail with Unrecognized node type: n8n-nodes-base.executeCommand. Remove it from NODES_EXCLUDE and restart n8n.
Is NODES_EXCLUDE="[]" safe?
It enables every node, including Local File Trigger. To bring back only Execute Command, set NODES_EXCLUDE='["n8n-nodes-base.localFileTrigger"]', which worked for us on n8n 2.46.
Can I use Execute Command on n8n Cloud?
No, n8n's docs say it isn't available on Cloud. Call ffmpeg over HTTP from an HTTP Request node instead; we tested that route on self-hosted n8n only.
The node is back but says ffmpeg: not found. Why?
The node runs commands in the shell where n8n runs, and that machine has no ffmpeg. Install it there; in the official Docker image, apk has to be restored first.
- n8nn8n Cloud has no Execute Command node and n8n 2.0 turns it off by default. Render a faceless short, burned-in captions or raw ffmpeg from one HTTP Request node, with the mp4 link back in the same node. Shorts $0.30 up to 60 s.
- Error: spawn ffmpeg ENOENTNode cannot find an ffmpeg program on PATH. Install it, spawn the ffmpeg-static path, or send the same args to an API.
- FileNotFoundError: No such file or directory: 'ffmpeg'Python asked the OS for a program called ffmpeg and found none on its PATH. Install it, use imageio-ffmpeg, or call ffmpeg over HTTP.
- ffmpeg exited with code 1: Conversion failed!fluent-ffmpeg reports only ffmpeg's exit code and last line. The real cause is in stderr: print it, or use -v error to see it alone.
- MakeMake has no module that runs ffmpeg and its HTTP module waits 300 s at most. Ask ReelSDK to hold the request for 280 s and the finished mp4 link comes back in the same module. Slideshows and captions $0.03 per output minute.
- ZapierZapier actions time out after 30 seconds and a Code step can't hold a video. Start a ReelSDK render from Webhooks by Zapier and catch the finished mp4 in a second Zap. Captions and slideshows $0.03 per output minute.
- Resize videoRender 9:16, 1:1, 4:5 and 16:9 copies of one video in a single ffmpeg call, about 2 cents for all 4 from a 30 s clip.
- ReelSDK vs self-hosted FFmpegTranscoding a minute of 1080p to 720p takes about 24 seconds of compute: about 0.8 cents on ReelSDK, about a tenth of a cent on a busy 4 vCPU box. Below about 11,000 minutes a month, the API costs less than an always-on box.
Ready to build?
Ship video from your app today
Your first $0.50 is free, and one key covers every endpoint.
