Write Cloud
Writes content from the message to a file in a connected cloud storage location.
Write Cloud takes a value from the message and stores it as a file in cloud storage.
Use a Write Cloud node in your Flow to persist something the Flow has produced: an export for another system to collect, a generated report, a copy of an incoming payload, or the result of a stream that has been collected back into a single value.
Write Cloud does not open a source stream and does not require Sink Collect. It works in an ordinary row.
To configure a Write Cloud node in the Flow Editor, drag it from the Stream group in the Palette to the Workspace canvas, then double-click the node to open and configure its behavior.
General tab
Name — optional display name for the node in the editor. By default the node shows a label of "Write Cloud"; the value you provide here replaces it.
Config — the tenant configuration. Defaults to the current tenant.
Connection — the storage connection to write to. Selected from the connections configured for the tenant. Required.
Remote Path — the path of the file to write. Required.
Input — the message property holding the content to write. Defaults to
msg.payload.Headers — an optional list of key and value pairs, added with the add button. Each row takes a header name and a value. The value is a typed input, so it can be entered directly as a string, number, boolean or JSON, or resolved from a msg, flow, global, or environment value, a JSONata expression, or a timestamp.
Remote paths
For AWS S3 or Azure Blob, if the connection does not already specify a bucket or container, Remote Path must begin with it, followed by the file path, using <bucket-or-container>/<path/to/file>. For example, test/hello.txt selects bucket or container test and file hello.txt.
For SFTP, Remote Path is passed through unchanged and must be a valid path on the remote server.
Writing to a path that already holds a file overwrites it in place. No error is raised and no confirmation is requested.
List Cloud fields vary by provider. Azure can return creation time, MD5, and content headers. AWS listings do not return createdAt or content headers and may omit contentMD5 for multipart objects. SFTP listings return only path, size, and updatedAt.
Preparing the content
Input is passed to the selected provider without serialization. Use a string, Buffer, or readable stream; serialize objects and other values before this node.
A Template node placed before Write Cloud is a convenient way to produce text, and unlike the Inject node it accepts multiple lines. Set its property to msg.payload, its format to Mustache template, and Output as to Plain text.
Text in a Mustache template is written literally, but {{ }} is a variable reference. Entering {{"hello world"}} looks for a variable of that name, finds nothing, and writes an empty value. To write the words themselves, enter them without braces.
Setting headers
Each Headers row pairs a name with a typed value, so a header can be fixed in the editor or drawn from the message at runtime.
For AWS S3 and Azure Blob, supported content headers such as contentType are normalized and applied. When contentType is omitted, the provider derives it from the file extension.
List Cloud output varies by provider. Azure includes available content properties. AWS list operations do not return content headers, and SFTP ignores Headers and returns only path, size, and updatedAt.
Source tab
The Source tab controls how the node behaves when it is processing a source stream rather than a single message.
Parallelism can be set to None, Ordered, or Unordered
Max — the concurrency ceiling for Ordered. Unordered can currently run one additional operation.
These are the same controls the Object nodes provide, such as Put Object and Query Object. A Write Cloud node placed inside a source stream can therefore write several files concurrently rather than one at a time.
The tab itself notes: Configurations in this tab are only active when this node is in a source. Outside a source stream the node processes a single message, and Parallelism and Max have no effect.
Example
A Flow writes a short text file to storage.
Prepare the content
Add a Template node with Property msg.payload, Format Mustache template, and Output as Plain text. Enter the text to write in the template body:
Configure the Write Cloud node
Set Connection to the storage connection, Remote Path to test/hello.txt, and leave Input as msg.payload.
Confirm the file was written
Request /write, then use a List Cloud node against the container. The new file appears with its path, size in bytes, and a contentType derived from the extension.
A Log node placed after Write Cloud shows the written content still on the message.
What the node returns
On success, Write Cloud preserves the configured Input value and passes the message downstream. It does not add a write result. In source mode, it updates the internal source pipeline.
Last updated
Was this helpful?