Jetformat

Where do my files go?

Jetformat runs on the machine that owns the source file and writes outputs to the path you provide.

By Jetformat

Office and standard PDF files stay on the machine

Scripts and agents call the local CLI. Office and standard PDF commands do not upload the source document to a remote conversion endpoint, and the conversion engine does not require Microsoft Office, WPS, or LibreOffice. Free commands run without a network connection. The only network traffic from a free command is the anonymous usage ping, which carries the command name and file type, never the file, and which jetformat telemetry off disables.

The online tools upload files

The online tools on jetformat.com are the exception: the files you choose there are uploaded and processed on Jetformat’s server. Uploads are deleted as soon as the job finishes, results 1 hour after the upload or when you press Delete now, and files are not used for training or viewed by people. When a document must not leave the machine, run the same operation with the CLI; every tool page shows the command.

Commands with separate service dependencies

The HTML-to-PDF beta sends the page to the configured renderer service. A renderer on another machine changes where that content is processed. The PDF timestamp command also contacts the chosen timestamp authority. Check those endpoints before applying a local-only policy to the entire CLI.

What billing sees

Paid commands report the verb, the number of units (pages or requests), a client-generated job id, and the CLI version. Paths, file names, file contents, and hashes are never sent. Failed jobs report zero.

Choose an explicit output

Use -o when a workflow needs a predictable destination:

jetformat word to-pdf contract.docx -o ./output/contract.pdf

After the command returns, the agent should report that exact path. Read the privacy page.

Review the processing boundary for your task

Local processing describes where the document operation runs. It does not mean every command has identical network behavior or that an agent host cannot access the text it receives. Consider the CLI, any configured service, and the application consuming the output separately.

Operation or integration Boundary to review
Office and ordinary PDF processing Local source and output paths
Optional usage telemetry Usage record described in the telemetry guide
Metered writes Account authorization and usage reporting
HTML capture Configured renderer and source resources
PDF timestamp Configured timestamp authority
Agent reading extracted text The agent host’s own handling of that text

Check an intended workflow before use

  1. Identify the exact commands, including any HTML capture or timestamp step.
  2. Review configured service addresses and determine which machines will receive inputs or requests.
  3. Choose local output paths with appropriate access for the intended recipient.
  4. Disable optional telemetry if that is required for the workflow, while accounting separately for paid-command authorization.
  5. Review logs and agent transcripts for source text or credentials before sharing them.

An output filename alone says nothing about who can access its directory or a later upload destination. Keep the original and derivative in locations appropriate to their contents, and share only the reviewed result. This guide describes the CLI’s boundaries; it does not certify the behavior of a separate agent platform or storage service.

Last updated