Cloudinary n8n integrationAutomate Cloudinary with n8n.

Sends binary file data from a previous node into the media library and returns the stored asset.

• File: the file to upload, required, usually the output of the node that fetched or generated it.
• Resource Type: image, video or raw, so a PDF or a ZIP is not treated as a picture.
• Additional Fields: Folder, Public ID, Tags, Structured Metadata as JSON, and Upload Preset if your account already defines one.

Cloudinary fetches the file itself from an address you pass in, so nothing has to be downloaded into the workflow first.

• URL: address of the image to upload, required, often {{ $json.url }} from the row or record being processed.
• Resource Type: pick video when the link points at a clip, otherwise image.
• Additional Fields: set Public ID to keep a stable name, Folder to sort, Tags to group.

Chains several transformation steps in order and returns one delivery URL plus a reusable transformation string.

• Public ID: required, the asset path without the file extension, for example folder/sample.
• Steps: each step picks an Action, Resize, Crop, Convert Format, Optimize, Trim or Raw Component, and they are applied in the order shown.
• Resource Type: image or video, which decides the URL path and the optimize output.

For teams that already know Cloudinary syntax, this builds the URL from a raw string used verbatim.

• Public ID: required, folder path included, extension left off.
• Transformation: required, the raw string where components are separated by a slash and qualifiers by commas, such as c_fill,g_auto,w_400,h_300/f_auto/q_auto.
• Format: an optional delivery extension without the leading dot.

Returns a delivery URL that serves the same picture in another file format.

• Public ID: required, identifies the asset to convert.
• Target Format: avif, webp, jpg, png or gif.
• Continue From Transformation: paste the transformation output of a previous Transform step to compound both into a single URL rather than starting fresh.

Cuts a picture down to fixed dimensions or to a ratio, and hands back the URL and the transformation string.

• Crop By: Dimensions for an exact Width and Height, or Aspect Ratio for a value like 16:9.
• Focus: Auto detects the important region, Face centers on a detected face, Center keeps the middle.
• Generative Fill: extends the image instead of cutting it, costs extra transformation credits and may render asynchronously on first request.

Auto format and auto quality in one step, which is the cheapest weight saving on any image you deliver.

• Public ID: required, the asset to optimize.
• Quality: Auto is the recommended setting, Best keeps more detail, Eco and Low trade quality for smaller files.
• Continue From Transformation: chain this after a Resize or Crop step, since neither of those adds optimization on its own.

Produces a delivery URL at the width and height you ask for, with a choice of how the picture fits the box.

• Width and Height: leave one at 0 to size by the other only.
• Fit: Limit never upscales, Fit may enlarge, the Pad modes keep the whole picture and fill the rest.
• Pad Background: Black by default, or Auto, Blurred, or a Background Color given as a name or a hex value.

Same framing logic as the image version, applied to a clip so it fits a different placement.

• Public ID: required, the video path without its extension.
• Crop By: Dimensions with Width and Height, or Aspect Ratio with a value such as 9:16.
• Focus: Auto, Face or Center decides what stays in frame once the edges are cut.

Lets Cloudinary choose the format, the codec and the quality for the browser that asks for the file.

• Quality: Auto for the balanced default, Good for a middle ground, Eco or Low when bandwidth matters more than detail.
• Continue From Transformation: wire the transformation field from a Trim or Resize step so the two compound into one URL.

Scales a clip to a target size and returns the URL to deliver it from.

• Width and Height: 0 on one side sizes by the other.
• Fit: Scale forces exact dimensions and may distort when both are set, the Pad modes letterbox instead.
• Pad Background: Auto samples a color from the clip, Blurred uses an enlarged copy, Solid Color takes the value you set.

Pulls one still frame out of a video and delivers it as an image.

• Frame: Auto lets Cloudinary pick a representative moment, Specific Time uses the Timestamp in seconds.
• Image Format: jpg, png, webp or gif for the generated still.
• Public ID: required, the source clip.

Shortens a clip through the delivery URL, leaving the stored file untouched.

• Start Offset and End Offset: times in seconds, and any combination of start, end and duration is accepted.
• Duration: clip length in seconds when you prefer to say how long rather than where to stop.
• Continue From Transformation: keeps an earlier crop or resize in front of the trim.

Removes one or several assets from the library, by public ID.

• Public ID(s): required, a single value, several separated by commas, or an expression returning an array.
• Type: required storage type, Upload, Private, Authenticated or Fetch.
• Delete Options: Keep Original deletes only the derived versions, Invalidate CDN clears cached copies, Transformations narrows the deletion to listed derived assets.

Reads the full record of a single asset so the rest of the workflow can branch on what it finds.

• Asset ID: required, the immutable asset_id, not the public ID.
• Options: Colors returns the predominant colors, Faces the detected face coordinates, Image Metadata the IPTC, XMP and EXIF values, Pages the page count of multi page files.

Queries the library with a Cloudinary search expression and returns the matching assets.

• Expression: for example folder:products/*, and an empty value matches everything.
• Resource Types: Cloudinary searches images only by default, so tick Video or Raw to widen it.
• Return All: enabled, n8n pages through every match; disabled, Max Results caps the answer between 1 and 500.
• With Field: add Tags or Context to avoid a second lookup downstream.

Writes structured metadata onto an asset identified by its asset ID.

• Asset ID: required, the immutable identifier returned by the library.
• Structured Metadata: required JSON, shaped like {"field1": "value1"}, matching the fields your account defines.
• Invalidate CDN: clears cached copies once the values change.

Sets or extends the tag list on one or more assets.

• Asset ID and Public ID(s): both required, because the tag append endpoint is keyed on public ID and does not accept an asset ID.
• Tags: required, a comma separated list of names.
• Mode: Set replaces the whole list, Append keeps what is already there.
• Type: must match the asset, since public IDs are scoped by resource type and type.

Builds the embed code and the configuration for the Cloudinary Video Player, returned as an embed URL and a player config string.

• Public ID: required, the clip to play.
• Source Types: fallback order, and HLS or MPEG-DASH cannot be combined with a Transformation that pins a format.
• Poster: the still shown before playback, ideally the secure URL of a Video: Thumbnail step.
• Playback: Muted, Loop, Autoplay Mode, and Generate Captions for AI captions in the spoken language.

Lists the structured metadata fields defined on the account, with no parameter to fill in.

Returns the tags in use for one kind of asset, which is how you check a label before applying it.

• Resource Type: required, image, video or raw.
• Prefix: keeps only the tags starting with the characters you give.
• Max Results: caps the list between 1 and 500.

The legacy route to the same result, kept for workflows built before the Asset resource existed. It is deprecated, and new work belongs on the Asset resource.

• Public ID: required, the asset to update, since this resource works by public ID rather than asset ID.
• Type: required storage type, Upload, Private, Authenticated or Fetch.
• Structured Metadata: required JSON of field keys and values.

Tagging through the deprecated public ID resource, useful only while an existing workflow has not been migrated.

• Public ID and Type: both required, they locate the asset in the older addressing scheme.
• Public ID(s): required, the assets to tag, single value or comma separated.
• Tags: required, comma separated names, applied with Mode set to Set or Append.