Introduction

What it does

Each of these is one cv command, or one job in the API.

• Convert a file to another format, such as a Word document to PDF, a deck to pictures or a workbook to CSV. There are 276 formats and 3,636 conversions between them.
• Read any file as Markdown or plain text, and a long one part by part. See Read a file.
• Write your own Markdown or HTML into a Word document, a PDF, a deck or an e-book. See Write a document.
• Edit cells of a spreadsheet, paragraphs of a Word document, text on a slide or fields of a PDF form, and keep the rest as it was. See Edit a file.
• Compare two versions of a file, part by part. See Compare two versions.

The CLI and the API

cv runs on your machine. It uploads the file, runs the job, waits for it and saves the result, so one command does what takes three requests in the API.

The API is HTTP and JSON at https://api.convert.dev/v1. Pick CLI, curl, Python or TypeScript with the icons at the top, or turn on --json for the CLI, and the examples on every page follow.

Agents learn every command from the skill file that cv carries. See Agent skill.

Files, jobs and conversions

You upload a file and start a job that names the format to make. The job runs one or more conversions and lists the files it made in outputs. Uploads and outputs are deleted after 24 hours.

Each conversion from one format to another is its own unit with a version, and only the conversions cv targets lists exist. Each one says what it keeps and what it drops, and a job's transforms names the conversions it ran and their versions.

Where to start

• Quickstart gets a key and converts a first file.
• Read a file and Edit a file are the guides most agents need.
• Jobs lists every field of a job in the API.
• Error codes says what each failure means and what to do next.